B. Parsons et al. The American Journal of Geriatric Pharmacotherapy

Gabapentin: A Pooled Analysis of Adverse Events from Three Clinical Trials in Patients with Postherpetic Neuralgia

Bruce Parsons, MD, PhD, Leslie Tire, PhD, and Sue Huang, PhD

Pfizer Inc., New York New York

ABSTRACT Background: Gabapcntin has bccn shown to bc well tolerated and effective in the management of the pain asso-

ciated with postherpctic neuralgia (PHN). It is assumed that adverse events occurring with gabapcntin arc dose related, their frequency and severity increasing with increasing doses.

Objective: The aim of this study was to assess the dose dependence of adverse events with gabapcntin by deter- mining the relationship between increasing doses of gabapentin and the onset a n d / o r worsening of adverse events in patients with PHN.

Methods: Data were pooled from 3 randomized, double-blind, placebo-controlled, parallel-group studies of gabapentin that focused on or included patients with PHN. Gabapcntin was initiated at 300 m g / d and titratcd to maintenance doses of 1800 to 3600 m g / d by day 12 to 24. The analysis of adverse events was based on 3 distinct groups: patients who received gabapcntin <1800 rag /d , those who received gabapcntin ---1800 rag /d , and those who received placebo. Patients who were given higher doses of gabapcntin had already received lower doses. An adverse event was recorded at the dose of its first onset and recorded again if its severity worsened at a higher dose.

Results: This study included data from 603 patients with PHN: 358 patients (196 [54.7%] women, 162 [45.3%] men; mean [SD] age, 72.3 [10.3] years) received gabapcntin, and 245 (133 [54.3%] women, 112 [45.7%] men; mean [SD] age, 73.3 [10.7] years) received placebo. The 3 most common adverse events were dizzi- ness, somnolence, and peripheral edema. Patients receiving gabapcntin ---1800 m g / d had a higher incidence of peripheral edema (7.5%) than those receiving gabapcntin <1800 m g / d (1.4%) or placebo (1.6'%) ( P < 0.002, gabapcntin ---1800 m g / d vs placebo). In contrast, the incidence of dizziness and somnolence was not higher in patients receiving gabapcntin -1800 m g / d compared with those in the other groups. Compared with placebo recipients, patients receiving gabapcntin <1800 m g / d reported a significantly greater frequency of dizziness (20.2'% gabapcntin <1800 m g / d vs 7.4% placebo; P < 0.002) and somnolence (14.9% vs 5.8%, respectively; 1" =

0.005). However, at ---1800 rag /d , rates of dizziness (9.7%) and somnolcncc (6.9'%) were comparable to thosc with placebo. Discontinuation rates wcrc comparable between patients receiving gabapentin and those receiving placebo.

Conclusions: In this pooled analysis of adverse-event data from 3 clinical trials in patients with PHN, the inci- dence of peripheral edema was increased when gabapcntin was titrated to -1800 mg /d . Dizziness and somnolence, the other most commonly occurring adverse events, were transient and did not occur 1Tlorc frequently or worsen with titration to -1800 mg /d . Based on these findings, it does not appear that safety concerns should limit titra- tion of gabapcntin to achieve optimal efficacy. ( A m ] Geriatr Pharmacother. 2004;2:157-162) Copyright © 2004 Excerpta Medica, Inc.

Key words: posthcrpetic neuralgia, gabapcntin, peripheral edema.

Accepted for publication July 18, 2004. Printed in the USA. Reproduction in whole or part is not permitted.

doi: I 0. 1016/j.amjopharm.2004,09.004 1543-5946/04/$19,00

Cop7right © 2_004 Excerpta b'ledica, Inc. Volume 2_ • Number 3 September 2004 15"I

The American Journal of Geriatric l'harmacotherapy B. l'arso~s et al.

I N T R O D U C T I O N For the majority of patients, the pain that accompanies herpes zoster infection (shingles) resolves sponta- neously as the skin lesions heal. However, in some patients, pain may persist even after the crusting of lesions. Postherpctic neuralgia (PHN) is a chronic, debilitating condition that occurs when pain localized to the affected dermatomes persists l~br ---3 months after the healing of the cutaneous manifestations of herpes zostcr) I t is estimated that >1 million new cases of her- pes zostcr infection occur each year and that -10% to 34% of these patients develop PHN. 14 This form of ncuropathic pain predominantly affects the elderly; nearly half o f patients aged ---60 years who arc infected with herpes zostcr develop PHN. s,6 Chronic debilitat- ing pain may cause depression, social isolation, sleep disturbance, and difficulty ambulating, possibly result- ing in falls, cognitive impairment, and malnutrition, 7 with consequent increases in health care costs.

Tricyclic antidepressants (eg, amitriptylinc, desipra- mine), opioids, topical lidocaine, and gabapcntin have all bccn effective in reducing the pain associated with P H N in randomized clinical trials. However, tricyclic antidepressants may produce undesirable anticholin- ergic adverse effects (cg, dry mouth, constipation, blurred vision), to which the elderly are particularly susceptible.8 M Moreover, postural hypotcnsion and cardiac arrhythmias are potential effects o f these agents, precluding their use in patients with underlying car- diovascular disease. Older anticonvulsants (cg, carba- mazcpinc, phenytoin) may cause potentially serious adverse effects (eg, hepatotoxicity, blood dyscrasias) that limit their usefulness. 12 16 Based on clinical experi- ence, the risk of physical or psychological dependence on opioids appears to be low when they arc used in the treatment of pain. Moreover, patients may develop tol- erance to some of the troublesome adverse effects o f opioids (cg, sedation, constipation, nausea, vomiting, itching), particularly when treatment is started at a low dose and titrated slowly upward. However, constipa- tion is usually persistent, and nausea and vomiting may require a switch to a different drug. The lidocaine patch is often well tolerated but may be associated with skin reactions. Given all these treatment options, fewer than half o f patients achieve relief o f PHN. 17 19 Finally, cognitive impairment, often an integral component of the emotional stress brought on by chronic pain, can be magnified in elderly patients when some of these agents are used. 2°

Because P H N primarily affects the elderly, the toler- ability of therapy is an important concern. I t is usually

assumed that adverse events are dose related, with the severity and frequency of events increasing with increasing doses. The present article presents the results o f an analysis o f the pooled results of 3 studies in patients with PHN. 21 2a The objective was to assess the dose dependence of adverse events with gabapentin by determining the relationship between titrated doses of gabapentin and the onset a n d / o r worsening of adverse events in patients with PHN.

P A T I E N T S A N D M E T H O D S The 3 clinical studies had similar designs: all were mult icentcr , randomized , double-bl ind, placebo- controlled, parallel-group trials. Two studies involved only patients with PHN, 21,22 whereas the other en- rolled patients with a variety of" neuropathic pain syn- dromes, 23 a l though only those with P H N were included in the present analyses. To bc included in the studies, patients had to be aged ---18 years and have at least moderate pain (defined as a score of---4 on an 11- point Likert pain scale in which 0 = no pain and 10 = worst possible pain). Furthermore, in the 2 P H N stud- its, 21,22 pain had to bc present for >3 months after healing of the herpes zoster rash. The primary efficacy measure was the change in the mean daily pain score from the baseline wcck to the final week.

In 2 studies, 21,22 patients wcrc randomly assigned to receive either gabapcntin or placebo. In the third study, 23 patients were randomly assigned to 1 of 3 treat- merit groups: gabapcntin 1800 m g / d , gabapcntin 2400 r ag /d , or placebo. After randomization in all studies, a titration period was fbllowcd by a stable-dose period l~br a total o f 7 or 8 weeks of treatment, depend- ing on the study design. The titration schedule varied slightly among the studies. In one 8-week stud); 21 the initial gabapentin dose of 300 mg/d was increased in a stcpwise manner over the 4-week titration period from 900, 1800, and 2400 m g / d to a maximum of 3600 m g / d administered in 3 divided doses. Doses were increased regardless of the efficacy response, unless intolerable adverse effkcts developed, in which case the dose could be decreased 1 level. The minimum dose permitted was 1200 m g / d . The dose established during the titration period was continued during the 4-week stable-dose period. In the 7-week study, 22 the gaba- pcntin dose was increased from 300 to 1200 111g/d dur- ing 4 days of l~brced titration and then to 1800 mg/d after 1 week. Patients randomized to receive 1800 m g / d remained at this dose; those randomized to receive 2400 m g / d were titratcd to this level after 2 weeks. Patients were discontinued from the study if they were

158

B. l'arso~ts et al. The American Journal of Geriatric l'harmacotherapy

unable to tolerate these fixed doses. In the other 8-week study, 23 the gabapcntin dose was titrated to 900 m g / d over 3 days and then increased over a S-week titration period to a maximum of 2400 r ag /d , depending on the cf'ficacy response.

Safkty assessments in each of the 3 studies included monitoring of vital signs, physical examinations, and adverse-event reports. The investigator questioned patients to elicit adverse events, defined as any unto- ward event that happened to the patient, whether or not it was considered serious or related to treatment. Case-report forms were used to record a description of the event, severity of the event, date started, date stopped, relationship to study drug, action taken, and outcome.

All patients with P H N who had fbllow-up infbrmation and received ->1 dose of study drug were included in an assessment of the 6 most common adverse events in the 3 pooled trials (dizziness, sonmolence, peripheral edema, diarrhea, nausea, and headache). Data analysis was perfbrmcd in 3 distinct groups: patients who were ran- domized to receive gabapentin <1800 rag /d , those who received gabapentin ->1800 r ag /d , and those who received placebo. Patients taldng higher doses of gabapentin had already received lower doses. Treatment- emergent adverse events were recorded at the dose of their first onset and recorded again if their severity wors- ened or they disappeared and re-emerged when the dose was increased.

Between-treatment comparisons (low-dose gabapentin vs placebo and high-dose gabapentin vs placebo) were peHbrmcd using analysis of covariancc. All P values

reported are 2-tailed, and no adjustments were made for multiple comparisons. Significance was set at P < 0.05.

R E S U L T S Study Population

Data were analyzed from 603 patients who were ran- domizcd to treatment and received ->1 dose of study drug: 229 patients from the study by Rowbotham et al, 21 334 from the study by Rice and Maton, 22 and 40 from the study by Scrpcll. 23 Three hundred fifty- eight patients (196 [54.7%] w o m e n , 162 [45.3%] men; mean [SD] age, 72.3 [10.3] years) received gabapentin, and 245 (133 [54.3%] WOlTlen, 112 [45.7%] men; mean [SD] age, 73.3 [10.7] years) received placebo. Eighty percent o f patients in both groups were aged ->65 years. The treatment groups were comparable with respect to sex, age, and race, although infbrmation on race was not available for over half o f patients.

Safety Analyses The safkty analyses included the following 3 groups:

patients who received gabapentin <1800 m g / d (n = 358), patients who received gabapcntin ->1800 m g / d (n = 321), and patients who received placebo (n = 245). The n l o s t conlnlon adverse events wcrc dizzi- ness, somnolence, and peripheral edema (Figure). O f these, dizziness was the most frequently reported, regardless of treatment assignment. However, patients receiving gabapentin <1800 m g / d reported dizziness significantly more often than those receiving placebo (20.2% vs 7.4%, respectively; P < 0.002). When the gabapentin dose was increased to ->1800 r ag /d , the

.fl £3-

~5 o~

2 5 -

2 0 -

15-

10-

5

0 - I Dizziness Somnolence Peripheral Edema

[ ] Low-dose gabapentin (n = 3s8)

[ ] High-dose gabapentin (n = 321)

• Placebo (n = 245)

Figure. Incidence of treatment-emergent adverse events reported at low gabapentin doses (< 1800 mg/d) and new or worsening adverse events reported at high gabapentin doses (->1800 mg/d) compared with placebo. *P < 0.002; P = 0.005; *P < 0.05, Fisher exact test.

159

The American Join'hal of Geriatric l'harmacotherapy B. l 'mvom et al.

incidence of new or worsening dizziness (9.7%) was con]parable to that ['or placebo.

Like dizziness, sonmolence was seen more frequently at lower doses: 14.9% of patients who received gaba- pentin <1800 m g / d reported somnolence, compared with 6.9% of patients reporting new or worsening som- nolence at higher gabapentin doses. The incidence of somnolence at lower doses was significantly greater than that with placebo (5.8%) (1 } < 0.001); however, there was no significant difference between higher doses and placebo (Figure).

The incidence of peripheral edema, in contrast to that of dizziness and somnolence, appeared to be dose related. There was a higher incidence of peripheral edema with gabapentin ---1800 m g / d con]pared with gabapentin <1800 m g / d (7.5% vs 1.4%, respectively). The incidence at the higher doses, but not the lower doses, was significantly greater than that with placebo (1.6%) (P < 0 .002)(Figurc) .

Adverse events wcrc also assessed at each dose of gabapcntin administered in the pooled studies (Tablc). The incidence of dizziness and somnolence increased as the dose was increased to 1800 m g / d (18.4% [59 /321] and 14.0% [4S/321] , respectively); however, the inci- dence decreased as the dose was increased to higher doses (7.7% [5 /65] and 9.2% [6 /65] at 3600 rag/d) . In contrast, the incidence of peripheral edema contin-

ued to increase gradually, with the highest incidence occurring at 3600 m g / d (12.3% [8 /65]) .

The incidence of the 3 other most common adverse events in patients receiving gabapentin (eg, diarrhea [16 /321 ; 5.0%], nausea [12 /321 ; 3.7%], headache [10 /321 ; 3.1%]) was too low to permit data analysis.

Discontinuation Rates 1Lates of discontinuation in the 3 studies were com-

parable between recipients of gabapentin and placebo. For example, in the study by Rice and Maton, 22 19.1% (22 /115) of patients randomized to receive gabapentin 1800 rag /d , 21.3'% (23 /108) receiving gabapcntin 2400 nag/d, and 15.3% (17 /111) receiving placebo discontinued treatment. Of these discontinuations, 34 of 45 (75.6%) patients receiving gabapentin and 7 of 17 (41.2%) patients receiving placebo discontinued due to adverse events, most occurring early in treatment. Among gabapentin recipients, 38.0% of withdrawals due to adverse events occurred within the first wcck of treatment and 76.0% within the first 3 weeks. The study by Rowbotham ct al 2] reported similar rates of discontinuation (21.2% [24 /113] of gabapentin recip- ients, 18.1% [21 /116] of placebo recipients). The pro- portion of patients discontinuing treatment fbr adverse events was comparable between the gabapcntin and placebo groups. Among patients achieving a stable dose

Table. Number (%) of patients with postherpetic neuralgia experiencing treatment-emergent adverse events, by gaba- pentin dose.

Dizziness Somnolence Peripheral Edema Dose, mg (No. of New New New Patients) Onset Onset Onset Continuing Worsening Continuing Worsening Continuing Worsening

300 (356) 18 (5, I) 0 0 15 (4,2) 0 600 (343) 32 (9,3) 14 (4,1) I (0.3) 26 (7,6) I I (3.2) 900 (347) 42 (I 2. I) 26 (7,5) 0 32 (9,2) 21 (6. I)

1200 (323) 51 (I 5.8) 31 (9,6) 0 35 (I 0.8) 24 (7.4) 1500 (308) 38 (I 2.3) 33 (I 0.7) 0 34 (I 1.0) 26 (8.4) 1800 (321) 59 (I 8.4) 38 (I 1.8) 0 45 (I 4.0) 28 (8.8) 2100 (183) 19 (10.4) 19 (10.4) 0 18 (9,8) 18 (9.8) 2400 (183) 23 (12.3) 14 (7,5) 2 (I.I) 21 (I 1.2) 16 (8.6) 2700 (74) 7 (9,5) 3 (4, I) 0 7 (9,5) 4 (5.4) 3000 (59) 2 (3,4) I (I ,7) 0 7 (11.9) 6 (I 0.2) 3300 (53) 3 (517) I (I ,9) 0 6 (11.3) 5 (9.4) 3600 (65) 5 (7,7) 2 (3, I) I (I .5) 6 (9,2) 4 (6.2)

0 0 0 0 0 I (0.3) 0 0 0 3 (0.9) I (0.3) 0 0 4 (i.2) 3 (0.9) 0 (0.3) 4 (i.3) 3 (i.o) 0 o 13 (4.0) 4 (I.2) 0 0 4 (2.2) 3 (i.6) 0 (0.5) 14 (7.5) 5 (2.7) 0 0 3 (4.1) 3 (4.1) 0 o 3 (s. i) 2 (3.4) o

0 3 (8.7) 2 (3.8) 0 (I .5) 8 (I 2.3) 4 (6.2) 0

Continuing adverse event continued from previous dose to current dose (previous dose is always smaller than current dose, although it is not necessar- ily the immediately preceding dose); worsening severit 7 of adverse event worse than at previous dose.

160

B. l'arsons et al. The American Journal of Geriatric l'harmaeotheral)y

o f gabapentin, 83.2% ( 9 4 / 1 1 3 ) received 2400 m g / d and 64.6% ( 7 3 / 1 1 3 ) received 3600 m g / d .

D I S C U S S I O N I t is evident t]-Ol-l-i the data in the table that fkwer patients received gabapcntin doses >1800 m g / d compared with gabapcntin doses ---1800 m g / d . One might conclude that this was because patients discontinued treatment at lower doses due to adverse effects. I t is important to consider that all patients receiving doses ---1800 m g / d first received 1800 m g / d . Furthermore, because o f the design o f the study by Pdcc and Maton, 22 115 patients wcrc randomized to receive gabapentin 1800 m g / d but wcrc never titrated to higher doses. This accounts in large part l~br the relative decrease in the total nmnbcr o f patients exposed to gabapcntin 2400 m g / d . Moreover, titration to 3600 lng/d was permitted only in the study by Rowbotham et al, 21 and 66.0% o f patients received that dose. Thus, the fhct that fkwcr patients received the higher doses is not primarily the consequence o f a sur- vivor ef/kct, as may occur when fkw patients arc exposed to higher doses because most have had difficuhy tolerat- ing a lower dose.

Additionally, data presented in the table indicate that the incidence o f dizziness and somnolence was lower at doses >1800 m g / d compared with 1800 lng/d. I t is difficuh to interpret these results precisely. The possibil- ity that patients experiencing these events at lower doses were no t titratcd to the higher doses is unlikely, as the gabapentin dose was increased beyond 1800 m g / d in most patients fbr w h o m titration was an option. A 1-nore likely explanation for the observed effects is that they were transient, became tolerable over time, and did no t worsen as the dose was increased to higher doses. Indeed, Pdce and Maton 22 and Scrpcll 2s reported that dizziness and somnolence were more c o m m o n early in the titration period.

Whereas dizziness and sonmolcncc did not appear to be dose-related effkcts, the incidence o f peripheral edema seemed to increase with increasing dose. The underlying etiology o f this edema remains unclear. I t was reported more frequently in adults than children and rarely resulted in discontinuation o f gabapentin treatment.

The d e v e l o p m e n t o f adverse events at doses < 1800 lng/d should not preclude titration to 1800 m g / d over a reasonable period o f time. Results fi 'om the 3 clinical trials discussed hcrc 21 23 support the conclusion that gabapcntin doses ->1800 m g / d arc no t associated with a higher incidence o f dizziness or somnolence, the 2 mos t COl-nl-non adverse events with gabapcntin. In fact, patients receiving gabapentin fbr several weeks had

a lower incidence o f dizziness and sonmolence than dur ing the titration phase, cvcn when the dose was increased to 3600 m g / d .

C O N C L U S I O N S In this pooled analysis o f adverse-event data frO1Yl 3 clinical trials in patients with P H N , the incidence o f peripheral edema was increased when gabapentin was titrated to ---1800 m g / d . Dizziness and sonmolence, the o ther mos t commonly occurr ing adverse events, were transient and did no t occur more frequently or worsen with titration to ---1800 m g / d . Based on these findings, it does no t appear that safety concerns should limit titration o f gabapentin to achieve optimal efficacy.

REFERENCES 1. Kost RG, Straus SE. Posthcrpctic ncur~gia pathogcncsis,

treatment and prevention. N E~lgl J Med. 1996;335:32 42.

2. Watson CP, Tyler KL, Bickers DR, ct al. A randomized vehicle controlled trial of topical capsaicin in the treat mcnt ofposthcrpctic neuralgia. Clin Tber. 1993;15:510 526.

3. Hcrnc K, Circlli R, Lcc P, Tyring SK. Antiviral therapy of acute herpes zostcr in older patients. Drugs Aging. 1996; 8:97 112.

4. Dworldn llM, Portcnoy RIC Pain and its persistence in herpes zostcr. Pabl. 1996;67:241 251.

5. dcMoragas JM, Kicrland RR. The outcome of patients with herpes zostcr. Arch DermatoL 1957;75:193 196.

6. Volmink J, Lancaster T, Gray S, Silagy C. Treatments for posthcrpctic neuralgia a systematic review of random izcd controlled trials. Faro Pract. 1996;13:84 91.

7. Fcrrcll BA. Overview of aging and pain. In: Fcrrcll BA, Fcrrcll BR, cds. Pain i~l the Elderly. 1st cd. Scattlc, Wash: IASP Press; 1996:1 10.

8. Max MB, Schafcr SC, Culnanc M, ct al. Amitriptylinc, but not lorazcpam, relieves posthcrpctic neuralgia. Neurology. 1988;38:1427 1432.

9. Kishorc Kumar R, Max MB, Schafkr SC, ct al. Dcsip raminc rclicvcs posthcrpctic neuralgia. Clin Pbarmacol

Ther. 1990;47:305 312. 10. Collins SL, Moore 1LA, McQuay HJ, Wiftkn P. Anti

dcprcssants and anticonvulsants tbr diabetic ncuropathy and posthcrpctic neuralgia: A quantitative systematic rcvicw. J Pain Symptom Ma~lwe. 2000;20:449458.

11. Watson CP, Evans RJ, Rccd K, ct al. Amitriptylinc versus placebo in post herpetic neuralgia. Neurology. 1982;32: 671 673.

12. McClcanc GJ. Intravenous infilsion of phcnytoin relieves ncuropathic pain: A randomized, double blinded,

161

The American Journal of Geriatric Pharmacotherapy B. Parsons et al.

placebo-controlled, crossover study. Anesth Analg. 1999; 89:985-988.

13. Rockliff BW, Davis EHH. Controlled sequential trials of carbamazepine in trigeminal neuralgia. Arch Neurol. 1966; 15:129-136.

14. Rull JA, Quibrera R, Gonzales-Millan H, Lozano Castaneda O. Symptomatic treatment of peripheral dia- betic neuropathy with carbamazepine (Tegretol): Double blind crossover trial. Diabetologia. 1969;5:215-218.

15. Spillane JD. The treatment of trigeminal neuralgia: Preliminary experience with Tegretol. Practitioner. 1964; 192:71-77.

16. Swerdlow M. The treatment of "shooting" pain. Postgrad Med J. 1980;56:159-161.

17. Sindrup SH, Jensen TS. Efficacy of pharmacological treat- ments of neuropathic pain: An update and effect related to mechanism of drug action. Pain. 1999;83:389-400.

18. Rowbotham MC, Reisner-I(eller LA, Fields HL. Both intravenous lidocaine and morphine reduce the pain of postherpetic neuralgia. Neurology. 1991;41:1024-1028.

19. Cherny NI, Thaler HT, Friedlander-Klar H, et al. Opioid responsiveness of cancer pain syndromes caused by neu- ropathic or nociceptive mechanisms: A combined analysis of controlled, single-dose studies. Neurology. 1994;44: 857-861.

20. Hart RP, Wade JB, Martelli MF. Cognitive impairment in patients with chronic pain: The significance of stress. Curr Pain Headache Rep. 2003;7:116-126.

21. Rowbotham M, Harden N, Stacey B, et al. Gaba- pentin for the treatment of postherpetic neuralgia: A randomized controlled trial. JAMA. 1998;280:1837- 1842.

22. Rice AS, Maton S, for the Postherpetic Neuralgia Study Group. Gabapentin in postherpetic neuralgia: A ran- domised, double blind, placebo controlled study. Pain. 2001;94:215-224.

23. Serpell MG, for the Neuropathic Pain Study Group. Gabapentin in neuropathic pain syndromes: A ran- domised, double-blind, placebo-controlled trial. Pain. 2002;99:557-566.

Address correspondence to: Bruce Parsons, MD, PhD, Pfizer Inc., 235 East 42nd Street, New York, NY 10017-5755.

162