dor neuripática

7/25/2019 Dor Neuriptica

1/10

Economic Evaluation of Oral Treatments for Neuropathic Pain

M. Soledad Cepeda* and John T. Farrar

*Department of Anesthesia and Clinical Epidemiology Unit, Javeriana University School of Medicine, Bogota,Colombia.Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania.

Abstract: The effectiveness of amitriptyline, carbamazepine, gabapentin, and tramadol for thetreatment of neuropathic pain has been demonstrated, but it is unknown which one is the most

cost-effective. We designed a cost-utility analysis of a hypothetical cohort with neuropathic pain of

postherpetic or diabetic origin. The perspective of the economic evaluation was that of a third-party

payor. For effectiveness and safety estimates, we performed a systematic review of the literature. For

direct cost estimates, we used average wholesale prices, and the American Medicare and Clinical

Laboratory Fee Schedules. For utilities of health states, we used the Health Utilities Index. We

modeled 1 month of therapy. For comparisons among treatments, we estimated incremental cost per

utility gained. To allow for uncertainty from variations in drug effectiveness, safety, and amount of

medication needed, we conducted a probabilistic Monte Carlo simulation. Amitriptyline was the

cheapest strategy, followed by carbamazepine, and both were equally beneficial. Gabapentin was

the most expensive as well as the least beneficial. A multivariable probabilistic simulation produced

similar results to the base-case scenario. In summary, amitriptyline and carbamazepine are more

cost-effective than tramadol and gabapentin and should be considered as first-line treatment for

neuropathic pain in patients free of renal or cardiovascular disease.

Perspective:Prescription practices should be based on the best available evidence, which includesthe evaluation of the medications cost-effectiveness. This does not mean that the cheapest or the

most expensive, but rather the most cost-effective medication should be chosenthe one whose

benefits are worth the harms and costs. We report a cost-effectiveness evaluation of treatments forneuropathic pain.

2006 by the American Pain Society

Key words:Neuropathic pain, cost-utility, amitriptyline, carbamazepine, gabapentin, tramadol.

Neuropathic pain is common in clinical practice. Pa-tients with conditions as diverse as diabetes, mul-tiple sclerosis, human immunodeficiency, or

stroke could develop pain of neuropathic origin. It isestimated that in the United States, there are more than3 million people with painful diabetic neuropathy and as

many as 1 million with postherpetic neuralgia.19

Pain inthese neuropathic pain syndromes greatly impairs all as-pects of quality of life.41,42

Treatment of pain in these neuropathic pain syn-

dromes improves quality of life.33 The effectiveness ofamitriptyline, carbamazepine, gabapentin, and tram-adol for the treatment of neuropathic pain has beendemonstrated in randomized double-blind controlledtrials and meta-analyses. However, there are no eco-nomic evaluations that assess these treatments, and it is

unknown which one is most cost-effective.Integrative studies that appraise the costs and benefits

(effectiveness and safety) of different strategies are cru-cial to guide clinicians in selecting the most cost-effectivetreatment.15,17,20,46Following is an economic evaluationof oral treatments for neuropathic pain.

Materials and Methods

Study DesignThis study was approved by the Institutional Review

Board of the Javeriana University School of Medicine,Bogota, Colombia. We designed a cost-utility analysis.

Received June 1, 2005; Revised September 1, 2005 and September 16,2005; Accepted September 18, 2005.

The Colombian Chapter of International Association for the Studyof Painprovided a grant that contributed to this research.

Address reprint requests to M. Soledad Cepeda, MD,PhD, Department ofAnesthesia andClinicalEpidemiologyUnit, JaverianaUniversity School ofMedicine, Cra 7 #40-62 Bogota, Colombia. E-mail: [email protected]

1526-5900/$32.00

2006 by the American Pain Society

doi:10.1016/j.jpain.2005.09.004

The Journal of Pain, Vol 7, No 2 (February), 2006: pp 119-128Available online at www.sciencedirect.com

119


2/10

Utilities represent the human preference for a specifichealth state. Utilities are assessed relative to 2 extremes:death, assigned a value of 0, and perfect health, assigneda value of 1. Utilities permit the combination of variousdisparate outcomes (ie, effectiveness and side effects ofan intervention) into a single composite summary out-come, providing an estimate of the broad picture of anintervention.16,17

We developed a decision analysis framework (Fig 1). Adecision analysis is the application of an explicit quanti-tative method for estimating financial costs and clinicaloutcomes of various strategies under conditions of un-

certainty, allowing the identification of the optimalstrategies.51 The decision tree for the decision analysiswas implemented using TreeAge Pro software (TreeAgeSoftware, Williamstown, Mass).

Target PopulationThe target population included patients with pain

from postherpetic neuralgia or diabetic neuropathy whowere free of cardiovascular, hepatic, and renal disease.

InterventionsWe evaluated amitriptyline, carbamazepine, tram-

adol, and gabapentin. The average and the range of

doses evaluated were extracted from clinical trials andpublished expert consensus19 (Table 1).

PerspectiveThe perspective of the economic evaluation was that

of a third-party healthcare payor.

Time HorizonWe modeled 1 month of therapy. This decision was

based on the facts that the pain relief and side effectsproduced by the studied medications have a rapid onsetand that in clinical practice if pain relief is not achievedor if major side effects develop within a short time the

medication is suspended. In addition, the lack of evi-

dence supporting the effectiveness of a combined ther-apy after treatment failure precluded us from modelinga longer time horizon.19

Data Sources

Effectiveness and Safety Estimation

For effectiveness and safety information, we per-formed a systematic review of the literature. Wesearched MEDLINE and the Cochrane Library for System-atic Reviews. Initially, the search was limited to system-atic reviews of the literature. Once the systematic re-views were identified, the search was expanded toinclude the Cochrane Central Register of Controlled Tri-als database to find randomized double-blind clinical tri-als published after the latest search reported in themeta-analyses. We excluded studies that were open orsingle masked, or that evaluated combined therapiessuch as amitriptyline with carbamazepine. There was nolanguage restriction.

The methodologic quality of the studies included wasassessed with the Oxford Scale.31 This scale evaluates 3items: information on the randomization procedure,quality of blinding, and the report of withdrawals. Themaximum score is 5 and the minimum is 1.

In order to combine the study results, we extracted thenumber of patients who had a clinically important levelof improvement (at least 50% pain relief)11,21,43 andcalculated the probability. If there were randomizeddouble-blind studies not included in the meta-analyses,we added these studies and updated the probability us-ing the reciprocal of the squared standard error of eachstudy as weights. These weights gave more importance

to studies with larger sample sizes. We also estimated the95% confidence intervals of the probabilities. In addi-tion, we reportthe Pvalue of the Q statistic used to assessthe homogeneity of the results;Pvalues smaller than .05are considered indicative that the studies are not homo-geneous.

For medication safety information, we followed thesame procedure described earlier to obtain the probabil-ities of having major and minor side effects. Major sideeffects were defined as resulting in the need to stop themedication owing to side effects or a determination of arisk of myocardial infarction.

Because randomized controlled trials (RCTs) are not

the best source for evaluating side effects (studies areusually powered to evaluate effectiveness, not side ef-

Figure 1. Decision model. The base-case is an individual withpain from postherpetic neuralgia or diabetic neuropathy. Theclinician may treat the patient with amitriptyline, carbamaz-epine, gabapentin, or tramadol. Patients could obtain pain re-lief and/or could develop major or minor side effects with thetreatment. Only the carbamazepine arm is shown; the otherstrategies are similar with the exception of developing myocar-dial infarction associated with the amitriptyline arm.

Table 1.Doses of the Medications Evaluatedin the Study

AMITRIPTYLINE CARBAMAZEPINE GABAPENTIN TRAMADOL

Average dose

(mg/day)

75 800 2400 200

Range

evaluated

(mg/day)

25200 6002400 9003600 100400

120 Economic Evaluation of Treatments for Neuropathic Pain


3/10

fects), we included additional nonrandom studies thatevaluated safety.

If the information for pain relief or side effects was notavailable, or was reported in terms of number needed totreat or to harm as opposed to number of patients withthe event, we contacted the corresponding authors toobtain the information.

Utility Estimation

For the utility of pain as a health state, we used the

Health Utilities Index Mark 3.6 These indices are generic

preference-scored systems for measuring health statusand health-related quality of life and for producing util-ity scores that represent community preferences. Theseindices have been developed by researchers at McMasterUniversity, Toronto, Ontario, during the past 20 years;they are reliable and have been validated extensively.6,22

The utilities for myocardial infarction and minor side ef-fects were obtained from studies that elicited the pref-erences directly from patients.34,58

For the assignment of utilities, we assumed, first, that

pain relief equated to a utility of 1; second, that minor

Table 2.Characteristics and Quality Score of Studies that Evaluate Effectiveness and Safety

STUDY TYPE OFSTUDY PATIENTS INTERVENTIONS QUALITYSCORE

Graff-Radford26 Randomized double-blind study Postherpetic neuralgia Amitriptyline range 12.5200

mg or placebo

4

Max35 Randomized double-blind

crossover study

Diabetic neuropathy Amitriptyline range 25150

or placebo

4

Max37

Randomized double-blindcrossover study

Postherpetic neuralgia Amitriptyline range 12.5150mg or placebo

3

Max36 Randomized double-blind

crossover study

Diabetic neuropathy Amitriptyline mean dose 105

mg or desipramine mean

dose 111 mg

3

Morello45 Randomized double-blind

crossover study

Diabetic neuropathy Amitriptyline range 2575

mg or gabapentin 900

1800 mg

4

Turkington59 Randomized double-blind study Diabetic neuropathy Amitriptyline 100 mg or

imipramine 100 mg

2

Vrethem61 Randomized double-blind

crossover study

Diabetic neuropathy Amitriptyline 75 mg or

placebo

4

Watson63 Randomized double-blind

crossover study

Postherpetic neuralgia Amitriptyline range 25137.5

mg or placebo

2

Watson

62

Randomized double-blindcrossover study Postherpetic neuralgia Amitriptyline mean dose 100mg range 37.5150 mg or

maprotiline

4

Watson64 Randomized double-blind Postherpetic neuralgia Amitriptyline 10150 mg or 5

study nortriptyline 50150 mg

Rull53 Randomized double-blind

crossover study

Diabetic neuropathy Carbamazepine 200600 mg

or placebo

4

Wiffen65 and Mcquay38

(including Rulls

study)

Systematic reviews of

randomized double-blind

studies

Trigeminal neuralgia Carbamazepine Excellent quality

Backonja8 Randomized double-blind study Diabetic neuropathy Gabapentin up to 3600 mg

or placebo

5

Gorson25 Randomized double-blind

crossover study

Diabetic neuropathy Gabapentin 900 mg or

placebo

3

Morello45 Randomized double-blind

crossover study

Diabetic neuropathy Gabapentin 9001800 mg or

amitriptyline 2575 mg

4

Reckless9 Randomized double-blind study Diabetic neuropathy Gabapentin 600, 1200, or

2400 mg or placebo

Unpublished trial

Rice50 Randomized double-blind study Postherpetic neuralgia Gabapentin 2400 mg or

placebo

5

Rice50 Randomized double-blind study Postherpetic neuralgia Gabapentin 1800 mg or

placebo

5

Rowbotham52 Randomized double-blind study Postherpetic neuralgia Gabapentin up to 3600 mg 5

Boureau10 Randomized double-blind study Postherpetic neuralgia Tramadol mean dose 275 mg

or placebo

5

Harati29 Randomized double-blind study Diabetic neuropathy Tramadol mean dose 210 mg

or placebo

5

Sindrup56 Randomized double-blind

crossover study

Diabetic neuropathy Tramadol 200400 mg or

placebo

4

121ORIGINAL REPORT/Cepeda and Farrar


4/10

side effects persisted throughout the 1-month study pe-riod; and third, that patients with major side effects sus-pended the treatment, and, therefore, did not obtainpain relief. We considered that the disutility of healthstates related to minor side effects was the same for alltypes of minor effects (eg, nausea, dizziness, dry mouth,or somnolence). We based this assumption on the factthat similar disutilities for these symptoms have beenreported in studies that have elicited utilities or prefer-ences in patients with cancer and nausea27,28 and in pa-tients with dizziness.23,47-49

Although it is common in economic evaluations to usethe utilities as weights to obtain quality-adjusted lifeyears, we did not consider this outcome because the timehorizon of this analysis wass limited to one month. Thus,we are using the utilities themselves as outcomes.

Cost Estimation

Only direct costs that were different among the treat-ment strategies were considered. For example, the diag-nostic tests for patients with diabetic neuropathy werenot included, because the costs of these tests are similarin all treatment strategies. For cost estimates of the med-ications, we obtained the average wholesale prices in the

Unites States in the year 2004. This information was ab-stracted from the Red Book.12 It was assumed that pa-tients received a 1-month prescription of medication.

For costs of minor side effects, we assumed that treat-ment was not required. With the exception of myocar-dial infarction, major side effects associated with thestudied medications are also transient, are very rarelylife-threatening, and subside when the medication isstopped. As such, we assumed that no specific treatmentwas required, but that an extra office visit to the physi-cian was necessary to seek another therapeutic option.All these assumptions reflect clinical practice.

The cost of a general practitioners office visit was ob-

tained from the 2004 American Medicare Fee Schedule.3

The cost of inpatient and outpatient care for an episode ofacute myocardial infarction was obtained from a previousstudy57 that evaluated cardiovascular events associatedwith the use of cyclooxygenase-2selective inhibitors in themanagement of chronic arthritis; the 2002 costs were con-verted to 2004 costs by assuming 4% inflation.5 The re-sources consumed included inpatient and outpatient care:cardiology Medicare diagnosis-related grouping (DRG) foruncomplicated myocardial infarction fee, emergency de-partment fee, initial cardiology consultation, interpreta-tion fee for echocardiography, angiographers fee for per-

cutaneous transluminal coronary angioplasty, follow-upvisit by cardiologist, and outpatient medication such as an-giotensin-converting enzyme inhibitor, beta-blocker, andaspirin, cardiac rehabilitation, and cardiologist office visit.

Because it is recommended that patients started oncarbamazepine have a complete blood count, includingplatelets, we included the cost of this test in the model,as obtained from the Clinical Laboratory Fee Schedule ofthe American Medical Association for 2004.4

Outcome MeasuresFor comparisons among treatments, we first calculated

the ratio of costs and utilities for each treatment strategyand then estimated the incremental cost per utilitygained. Incremental ratios were obtained as the differ-ence between costs divided by the difference in utilitiesbetween strategies. This ratio represents the extra ben-efit per extra dollar spent. Therefore, incremental ratioswere estimated only for strategies that were more effec-tive than the cheapest strategy. No discount for benefitsor costs was applied, because the time span of the eco-nomic evaluation was 1 month.24

Sensitivity AnalysesTo allow for uncertainty from variations in drug effec-

tiveness, safety, amount of medication needed to

Table 3.Amitriptyline Clinical Probability Estimates

STUDY

PROBABILITY OF

HAVINGPAIN

RELIEF

PROBABILITY OFHAVING

MINORSIDEEFFECTS


MAJORSIDEEFFECTS

Graff-Radford26 Not reported Not reported 1/11

Max35 15/29 28/29 3/37

Max37

15/34 30/34 5/34Max36 28/38 24/38 7/38

Morello45 14/21 15/21 2/21

Turkington59 19/20 Not reported Not reported

Vrethem61 5/19 3/19 1/19

Watson63 16/24 16/24 1/24

Watson62 15/32 20/35 3/35

Watson64 21/31 28/33 1/33

Base-case weighted probabilities .68 .80 .07

Test of homogeneity .0001 .0001 .46

Base-case probability of having myocardial .001713

infarction (sensitivity analysis) (.001.01)

Probabilities used in the sensitivity

analyses

.57.71 .8.85 .06.09



5/10

achieve pain relief, and health state preference for myo-cardial infarction, we conducted a probabilistic MonteCarlo simulation (1,000 trials), assuming that the vari-ables had a triangular distribution.57

The probabilities evaluated in the sensitivity analyseswere obtained by excluding studies that were outliers.For costs, we used the range of plausible costs associatedwith the different doses of medication reported in thestudies, medical office visits, and inpatient and outpa-tient care for an acute myocardial infarction. To incorpo-rate uncertainty of myocardial infarction utility, we useda score of .8 which gave a greater disutility than thelower limit of the 95% confidence interval for the myo-cardial infarction utility score, which was .84.58

Results

The searches were conducted of data from 1966 to thethird week of June 2004. We used the following MedicalSubject Headings (MeSH) terms: neuropathic pain; re-view literature or randomized controlled trials or meta-

analysis; diabetic neuropathies or peripheral nervous sys-tem diseases; pain or neuralgia or nervous systemdiseases; tramadol; anticonvulsants, or gabapentin; anti-depressive agents, tricyclic or amitriptyline; carbamaz-epine.

We located 12 meta analyses2,7,9,14,18,32,38-40,55,60,65

that evaluated amitriptyline, carbamazepine, tramadol,and gabapentin for postherpetic neuralgia and/or dia-betic neuropathy. In addition, we found 4 randomizeddouble-blind trials that evaluated amitriptyline,26,45,59,64

2 that evaluated gabapentin,45,54 and 1 that evaluated

tramadol10 that were not included in the previous meta-

analyses and were included in our analysis.

We estimated the risk of myocardial infarction from a

pharmacoepidemiologic study that evaluated the excess

risk of having myocardial infarction.13

Clinical Effectiveness and SafetyOnly 2 of the 5 RCTs evaluated carbamazepine related

to the syndromes we were studying. One was a crossover

trial that evaluated diabetic neuropathy,65 and the other

evaluated other pain syndromes in addition to diabetic

neuropathy and postherpetic neuralgia and therefore

had to be excluded.30 Because carbamazepine is com-

monly used in clinical practice, we decided to employ the

pooled estimates for effectiveness and safety from the

latest systematic review of literature that assessed car-

bamazepine in trigeminal neuralgia,65 which is also a

neuropathic pain syndrome. It is worth noticing that the

safety and effectiveness estimates reported in the othermeta-analyses14,38,55 that evaluated carbamazepine in

trigeminal neuralgia and the estimate from the 1 trial

that evaluated diabetic neuropathy were all similar.

A description of the characteristics and quality score of

the studies that assessed effectiveness and safety are dis-

played inTable 2.The median quality scores of the stud-

ies that evaluated amitriptyline, carbamazepine, gabap-

entin, and tramadol were 4, 4, 4.5, and 5 respectively.

The probabilities of having 50% or more of pain relief,

Table 4.Carbamazepine Clinical Probability Estimates

STUDY

PROBABILITY OF

HAVINGPAINRELIEF


MINORSIDEEFFECTS


MAJORSIDEEFFECTS

Rull53 28/30 16/30 2/30

Wiffen65 (including Rulls study)* 234/383 101/230 10/194


Probabilities used in the sensitivityanalyses

.57.67 .44.53 .05.10

*Total of 4 studies.

Table 5.

Gabapentin Clinical Probability Estimates

STUDY

PROBABILITY OF

HAVINGPAINRELIEF


MINORSIDEEFFECTS


MAJORSIDEEFFECTS

Backonja8 47/82 71/84 7/84

Gorson25 14/19 12/19 0/19

Morello45 11/21 16/21 2/21

Reckless9 70/244 Not reported 24/244

Rice (2400 mg)50 42/98 62/108 19/108

Rice (1800 mg)50 44/107 66/115 15/115

Rowbotham52 47/109 62/113 15/113


Test of homogeneity .0001 .0001 .11

Probabilities used in the sensitivity

analyses

.47.52 .5.65 .08.1



6/10

major side effects, or minor side effects associated witheach treatment are shown inTables 3to 6.

Utility AssessmentThe utility of having persistent pain was set at 0.55.

This utility score corresponds in the utilities indices tomoderate pain that prevents some activities. The util-

ity of obtaining pain relief without side effects was set at1. If there were minor side effects, the utility was set at0.95.34 The utility of myocardial infarction was set at0.88.57,58

Cost EstimatesThe costs of the medications, medical office visits, in-

patient and outpatient care for an acute myocardial in-farction, and laboratory tests are shown inTable 7.

Cost-Utility AnalysisAmitriptyline was the cheapest strategy, followed by

carbamazepine, and both were equally beneficial. Tram-adol and gabapentin were dominated strategies, mean-ing that these strategies were not only more expensive,but that they also produced less benefit than the mosteffective therapies (seeTable 8).

Sensitivity AnalysesThe plausible values used in the sensitivity analyses can

be seen inTables 2to 7. The multivariable probabilistic

sensitivity analyses produced results similar to thebasecase scenario. Amitriptyline and carbamazepinewere more beneficial and less expensive than tramadoland gabapentin (Table 8). The distribution of costs andutilities in each of the trials is shown inFigure 2.

DiscussionThis economic evaluation shows that in patients free of

cardiovascular hepatic or renal disease, older treatmentsfor neuropathic pain, such as amitriptyline and carbam-azepine, are better options than the newer and moreexpensive treatments.

A major strength of this study is the focus on evidence-based information to populate the decision tree analysisand to provide the results. While the assumptions under-lying our evaluation can be questioned, they have beencarefully described and are based on the best availabledata. This allows other investigators the opportunity to

contribute additional information or repeat our analysiswith other sets of assumptions, as may be appropriateonce new information becomes available.

There are a number of limitations that also should beconsidered in any study of this type. In this study, painfuldiabetic neuropathy and postherpetic neuralgia wereconsidered as 1 entity because they are both examples ofneuropathic pain for which there are data from each ofthe agents. There is clinical consensus that the first-line

Table 6.Tramadol Clinical Probability Estimates

STUDY

PROBABILITY OF

HAVING PAINRELIEF


MINORSIDEEFFECTS


MAJORSIDEEFFECTS

Boureau10 40/63 13/64 6/64

Harati29 43/63 Not reported 9/63

Sindrup56 11/34 28/34 7/34

Base-case weighted probabilities .59 .43 .12Test of homogeneity .001 .001 .3

Probabilities used in the

sensitivity analyses

.5.66 .4.8 .12.16

Table 7.Cost Estimates

AMITRIPTYLINE CARBAMAZEPINE GABAPENTIN TRAMADOL

Average dose (mg/day) 75 800 2400 200

Range evaluated 25200 6002400 9003600 100400

Tablet (mg) 25 200 300 50

Cost per tablet 0.15 0.31 1.11 0.78

Cost of a month prescription 13.5 37.2 266.4 93.6

Range of cost of a month

prescription

4.536 27.9111.6 99.9399.6 46.8187.2

Cost of physician office visit 35 35 35 35

(Range evaluated) (2060) (2060) (2060) (2060)

Cost of complete blood

count test

10.86

Cost of in- and outpatient

care for an acute

myocardial infarction

12,642 (5,32422,000)



7/10

therapies, such as the ones we evaluated, produce similarresults in multiple types of neuropathic pain.19 This per-

ception is supported by clinical trials and meta-analysesin which the magnitude of the treatment effect of antide-pressants, anticonvulsants, and tramadol is similar in dia-betic neuropathy and postherpetic neuralgia.14,18,38,55,65

Combining these syndromes allowed us to obtain a moreprecise estimate of the treatment effect. Despite thecombination, only 1 study that evaluated carbamazepine

had the necessary information to be included in the anal-ysis; thus we decided to include studies that evaluated

trigeminal neuralgia. We believed that this inclusion wasreasonable because trigeminal neuralgia is also a classicneuropathic pain syndrome, and the treatment effectsize and side-effect rate were similar in the carbamaz-epine studies.

The quality of an economic evaluation depends on in-cluding all appropriate costs in the analysis. A major con-

Figure2. Cost-utility scatter plot. This figure shows the cost-utility ratio for amitriptyline, carbamazepine, gabapentin, and tramadolobtained after 1,000 simulation trials. The goals of the multivariable sensitivity analysis were to incorporate likely variations in costs,effectiveness, safety, and utility values and observe the effect of this uncertainty. Cost, effectiveness, safety, and utility scoreestimates were varied over the full range of plausible values. Amitriptyline (xs) was the cheapest strategy, followed by carbamaz-epine (solid squares), and both were equally beneficial. Gabapentin (solid triangles) isthe mostexpensive treatment and the onethatproduces the smallest benefit.

Table 8.Result of Cost-Utility Analysis Using the Base-Case Scenario and the ProbabilisticSensitivity Analyses

ANALYSIS STRATEGY

COST/PATIENT/

MONTH ($)

INCREMENTAL

COST($)

EFFECTIVENESS

(UTILITIES)

INCREMENTAL

EFFECTIVENESS

INCREMENTAL COST-

EFFECTIVENESSRATIO

Base-case scenario Amitriptyline 29 .807

Carbamazepine 50 20 .807 0 Dominated

Tramadol 98 68 .769

.038 DominatedGabapentin 270 241 .697 .11 Dominated

Multivariable

sensitivity

analysis

Amitriptyline 56.0 .797

Carbamazepine 72.5 16.5 .798 .0004 43.296

Tramadol 114.1 41.6 .761 .037 Dominated

Gabapentin 256.1 183.6 .737 .061 Dominated

NOTE. All options referenced to the cheapest strategy (amitriptyline).



8/10

cern in the use of amitriptyline is the documented in-creased risk of myocardial infarction. To overtly dealwith this, we incorporated this risk in the model, eventhough this association is of unclear importance in pa-tients free of cardiovascular disease.1 We excluded fromthe analysis the reported possibility of decreases in plate-let or white blood cell counts with the use of carbamaz-

epine, because there are no available data to estimateaccurately the incidence or repercussions of these veryrare events. In addition, most cases of leukopenia havenot progressed to the more serious conditions of aplasticanemia or agranulocytosis. In addition, decreases inwhite blood cell counts have also been described withgabapentin.66 Nonetheless, we included the costs of theblood tests routinely performed when carbamazepine isinitiated. The exclusion of very rare events should notthreaten the validity of our results.

We modeled 1 month of therapy. Although this deci-sion was based on the short follow-up of the studies andthe fact that the pain relief and side effects produced by

the studied medications have a rapid onset, in clinicalpractice patients need to be on the medication for alonger period of time. The generalization of the resultsto this setting is not necessarily straightforward; whereasside effects could decrease over time, effectiveness coulddecrease as well. Still, we considered that the results of astudy with a longer time horizon could become specula-tive owing to lack of validated data.

We assumed that if pain relief occurred, patients werein perfect health with a utility score value of 1; howeverin real life, patients may have other heath problems thatare not reflected in that utility score. Because this as-

sumption applies to all the strategies, it did not favor anystrategy in particular and did not bias the resultstherelative differences among strategies would remain thesame despite the use of a different utility score.

The quality of an economic evaluation also depends onthe quality of the search and data obtained. The searchfor the probabilities of having major or minor side ef-

fects or pain relief was systematic, comprehensive, andbased on numerous meta-analyses and recent random-ized double-blind controlled trials. However, the samplesize of the published studies, with the exception of gaba-pentin, was small, which tends to overestimate treat-ment effects size.44 This may have enhanced the appar-ent usefulness of the older agents in this analysis.

In the sensitivity analysis, we decreased the likelihoodof having pain relief and increased the probabilities ofhaving side effects by excluding studies with the highestor lowest estimates, and the results remained similar,with tramadol and gabapentin dominated by cheaperand more effective therapies. This favors our original

findings. However, we caution against the generaliza-tion of the findings of this study to patients with comor-bidites such as cardiovascular or renal disease in whichamitriptyline may be contraindicated. In the presence ofcardiovascular, hepatic, or renal disease, tramadol, gaba-pentin, or topical agents such as capsaicin should be con-sidered as the first line of treatment.

In summary, amitriptyline and carbamazepine aremore cost-effective than tramadol and gabapentin, andshould be considered as first-line treatments for neuro-pathic pain of diabetic or postherpetic origin in patientsfree of cardiovascular, hepatic, or renal disease.

References

1. Agency for Health Care Policy and Research. Treatmentof depressionnewer pharmacotherapies. Summary, evi-dence report. Technology assessment number 7. Availableat: http://www.ahcpr.gov/clinic/epcsums/deprsumm.htm.Accessed March 7, 2004

2. Gabapentin new indication. In postherpetic neuralgiawhen amitriptyline fails. Prescrire Int 11:111-112, 2002

3. Centers for Medicare & Medicaid Services. Available at:

http://www.cms.hhs.gov/physicians/mpfsapp/default.asp.Accessed March 7, 2004

4. Clinical laboratory fee schedule. Available at:http://www.cms.hhs.gov/providers/pufdownload/clfdown.asp. AccessedJuly 11, 2004

5. Consumer price indices. Available at: http://www.dmi.uiuc.edu/cp/glossary04/gcpi.htm.Accessed July 7, 2004

6. Health Utilities Group. Health utilities index and quality-of-life. Available at: http://www-fhs.mcmaster.ca/hug/.Ac-cessed July 7, 2004

7. Alper BS, Lewis PR: Treatment of postherpetic neuralgia:a systematic review of the literature. J Fam Pract 51:121-128,2002

8. Backonja M, Beydoun A, Edwards KR, Schwartz SL, Fon-

seca V, Hes M, LaMoreaux L, Garofalo E: Gabapentin for thesymptomatic treatment of painful neuropathy in patientswith diabetes mellitus: a randomized controlled trial. JAMA280:1831-1836, 1998

9. Backonja M, Glanzman RL: Gabapentin dosing for neuro-pathic pain: evidence from randomized, placebo-controlledclinical trials. Clin Ther 25:81-104, 2003

10. Boureau F, Legallicier P, Kabir-Ahmadi M: Tramadol inpost-herpetic neuralgia: a randomized, double-blind, placebo-controlled trial. Pain 104:323-331, 2003

11. Cepeda MS, Africano JM, Polo R, Alcala R, Carr DB: Whatis the change in the pain intensity levels that is meaningfulto the patients? Pain 105:151-157, 2003

12. Cohen HE: Red Book. 2003 Drug Topics. Montvale, NJ,Thompson PDR, 2003, pp 1-780

13. Cohen HW, Gibson G, Alderman MH: Excess risk of myo-cardial infarction in patients treated with antidepressantmedications: association with use of tricyclic agents. Am JMed 108:2-8, 2000

14. Collins SL, Moore RA, McQuay HJ, Wiffen P: Antidepres-sants and anticonvulsants for diabetic neuropathy and pos-therpetic neuralgia: a quantitative systematic review. J PainSymptom Manage 20:449-458, 2000

15. Detsky AS, Naglie IG: A clinicians guide to cost-effectivenessanalysis. Ann Intern Med 113:147-154, 1990

http://www.ahcpr.gov/clinic/epcsums/deprsumm.htmhttp://www.ahcpr.gov/clinic/epcsums/deprsumm.htmhttp://www.cms.hhs.gov/physicians/mpfsapp/default.asphttp://www.cms.hhs.gov/physicians/mpfsapp/default.asphttp://www.cms.hhs.gov/providers/pufdownload/clfdown.asphttp://www.cms.hhs.gov/providers/pufdownload/clfdown.asphttp://www.dmi.uiuc.edu/cp/glossary04/gcpi.htmhttp://www.dmi.uiuc.edu/cp/glossary04/gcpi.htmhttp://www.dmi.uiuc.edu/cp/glossary04/gcpi.htmhttp://www-fhs.mcmaster.ca/hug/http://www-fhs.mcmaster.ca/hug/http://www-fhs.mcmaster.ca/hug/http://www.dmi.uiuc.edu/cp/glossary04/gcpi.htmhttp://www.dmi.uiuc.edu/cp/glossary04/gcpi.htmhttp://www.cms.hhs.gov/providers/pufdownload/clfdown.asphttp://www.cms.hhs.gov/providers/pufdownload/clfdown.asphttp://www.cms.hhs.gov/physicians/mpfsapp/default.asphttp://www.cms.hhs.gov/physicians/mpfsapp/default.asphttp://www.ahcpr.gov/clinic/epcsums/deprsumm.htmhttp://www.ahcpr.gov/clinic/epcsums/deprsumm.htm


9/10

16. Drummond MF, OBrien BJ, Stoddart GL, Torrance GW:Cost-utility analysis, in Drummond MF, OBrien BJ, StoddartGL, Torrance GW (eds): Methods for the Economic Evalua-tion of Health Care Programmes. Oxford, Oxford UniversityPress, 1997, pp 139-204

17. Drummond MF, Richardson WS, OBrien BJ, Levine M,Heyland D: Users guides to the medical literature. XIII. Howto use an article on economic analysis of clinical practice. A.

Are the results of the study valid? Evidence-Based MedicineWorking Group. JAMA 277:1552-1557, 1997

18. Duhmke RM, Cornblath DD, Hollingshead JRF: Tramadolfor neuropathic pain. Cochrane Database Syst Rev 2, 2004,CD003726

19. Dworkin RH, Backonja M, Rowbotham MC, Allen RR, Ar-goff CR, Bennett GJ, Bushnell MC, Farrar JT, Galer BS, Hay-thornthwaite JA, Hewitt DJ, Loeser JD, Max MB, Saltarelli M,Schmader KE, Stein C, Thompson D, Turk DC, Wallace MS,Watkins LR, Weinstein SM: Advances in neuropathic pain:diagnosis, mechanisms, and treatment recommendations.Arch Neurol 60:1524-1534, 2003

20. Eisenberg JM: Clinical economics. A guide to the eco-nomic analysis of clinical practices. JAMA 262:2879-2886,

1989

21. Farrar JT, Young JP, LaMoreaux L, Werth JL, Poole RM:Clinical importance of changes in chronic pain intensitymeasured on an 11-point numerical pain rating scale. Pain94:149-158, 2001

22. Feeny DH, Torrance GW, Furlong WJ: Health utility in-dex, in Spilker B (ed): Quality of Life and Pharmacoeconom-ics in Clinical Trials. Philadelphia, PA, Lippincott-Raven,1996,pp 239-252

23. Fielder H, Denholm SW, Lyons RA, Fielder CP: Measure-ment of health status in patients with vertigo. Clin Otolar-yngol 21:124-126, 1996

24. Gold MR, Siegel JE, Ruseell LB, Weinstein MC: Theoreti-

cal foundations of cost-effectiveness analysis, in Gold MR,Siegel JE, Ruseell LB, Weinstein MC (eds): Cost-Effectivenessin Heath and Medicine. Oxford, Oxford University Press,1996, pp 25-53

25. Gorson KC, Schott C, Herman R, Ropper AH, Rand WM:Gabapentin in the treatment of painful diabetic neuropa-thy: a placebo controlled, double blind, crossover trial.J Neurol Neurosurg Psychiatry 66:251-252, 1999

26. Graff-Radford SB, Shaw LR, Naliboff BN: Amitriptylineand fluphenazine in the treatment of postherpetic neural-gia. Clin J Pain 16:188-192, 2000

27. Grunberg SM, Boutin N, Ireland A, Miner S, Silveira J,Ashikaga T: Impact of nausea/vomiting on quality of life as avisual analogue scale-derived utility score. Support Care

Cancer 4:435-439, 1996

28. Grunberg SM,Srivastava A, Grunberg KJ, Weeks J: Inten-sity of chemotherapy-induced emesis and overall survival asdeterminants of a global utility score. Support Care Cancer10:624-629, 2002

29. Harati Y, Gooch C, Swenson M, Edelman S, Greene D,Raskin P, Donofrio P, Cornblath D, Sachdeo R, Siu CO, KaminM: Double-blind randomized trial of tramadol for the treat-ment of the pain of diabetic neuropathy. Neurology 50:1842-1846, 1998

30. Harke H: The response of neuropathic pain and pain incomplex regional pain syndrome I to carbamazepine andsustained-release morphine in patients pretreated with spi-nal cord stimulation: a double-blinded randomized study.Anesth Analg 92:488-495, 2001

31. Jadad AR, Moore RA, Carroll D, Jenkinson C, ReynoldsDJ, Gavaghan DJ, McQuay HJ: Assessing the quality of re-ports of randomized clinical trials: is blinding necessary?Control Clin Trials 17:1-12, 1996

32. Joss JD: Tricyclic antidepressant use in diabetic neurop-athy. Ann Pharmacother 33:996-1000, 1999

33. Katz N: The impact of pain management on quality of

life. J Pain Symptom Manage 24(suppl 1):S38-S47, 200234. Lawrence WF, Fryback DG, Martin PA, Klein R, Klein BE:Health status and hypertension: a population-based study.J Clin Epidemiol 49:1239-1245, 1996

35. MaxMB, Culnane M, Schafer SC, Gracely RH,Walther DJ,Smoller B, Dubner R: Amitriptyline relieves diabetic neurop-athy pain in patients with normal or depressed mood. Neu-rology 37:589-596, 1987

36. Max MB, Lynch SA, Muir J, Shoaf SE, Smoller B, DubnerR: Effects of desipramine, amitriptyline, and fluoxetine onpain in diabetic neuropathy. N Engl J Med 326:1250-1256,1992

37. Max MB, Schafer SC, Culnane M, Smoller B, Dubner R,

Gracely RH: Amitriptyline, but not lorazepam, relieves pos-therpetic neuralgia. Neurology 38:1427-1432, 1988

38. McQuay H, Carroll D, Jadad AR, Wiffen P, Moore A:Anticonvulsant drugs for management of pain: a systematicreview. Br Med J 311:1047-1052, 1995

39. McQuay HJ, Tramer M, Nye BA, Carroll D, Wiffen PJ,Moore RA: A systematic review of antidepressants in neuro-pathic pain. Pain 2-3:68:217-227, 1996

40. Mellegers MA, Furlan AD, Mailis A: Gabapentin for neu-ropathic pain: systematic review of controlled and uncon-trolled literature. Clin J Pain 17:284-295, 2001

41. Meyer-Rosberg K, Burckhardt CS, Huizar K, KvarnstromA, Nordfors LO, Kristofferson A: A comparison of the SF-36

and Nottingham Health Profile in patients with chronic neu-ropathic pain. Eur J Pain 5:391-403, 2001

42. Meyer-Rosberg K, Kvarnstrom A, Kinnman E, Gordh T,Nordfors LO, Kristofferson A: Peripheral neuropathicpaina multidimensional burden for patients. Eur J Pain5:379-389, 2001

43. Moore A, Moore O, McQuay H, Gavaghan D: Derivingdichotomous outcome measures from continuous data inrandomised controlled trials of analgesics: use of pain inten-sity and visual analogue scales. Pain 69:311-315, 1997

44. Moore RA, Gavaghan D, Tramer MR, Collins SL, McQuayHJ: Size is everythinglarge amounts of information areneeded to overcome random effects in estimating directionand magnitude of treatment effects. Pain 78:209-216, 1998

45. Morello CM, Leckband SG, Stoner CP, Moorhouse DF,Sahagian GA: Randomized double-blind study comparingthe efficacy of gabapentin with amitriptyline on diabeticperipheral neuropathy pain. Arch Intern Med 159:1931-1937, 1999

46. OBrien BJ, Heyland D, Richardson WS, Levine M, Drum-mond MF: Users guides to the medical literature. XIII. Howto use an article on economic analysis of clinical practice. B.What are the results and will they help me in caring for mypatients? Evidence-Based Medicine Working Group. JAMA277:1802-1806, 1997

47. Perez N, Garmendia I, Garcia-Granero M, Martin E, Garcia-Tapia R: Factor analysis and correlation between DizzinessHandicap Inventory and Dizziness Characteristics and Impacton Quality of Life scales. Acta Otolaryngol 545:145-154, 2001



10/10

48. Perez N, Martin E, Garcia-Tapia R: Dizziness: relating theseverity of vertigo to the degree of handicap by measuringvestibular impairment. Otolaryngol Head Neck Surg128:372-381, 2003

49. Prieto L, Santed R, Cobo E, Alonso J: A new measure forassessing the health-related quality of life of patients withvertigo, dizziness or imbalance: the VDI questionnaire. QualLife Res 8:131-139, 1999

50. Rice AS, Maton S, Postherpetic Neuralgia Study Group:Gabapentin in postherpetic neuralgia: a randomised, dou-ble blind, placebo controlled study. Pain 94:215-224, 2001

51. Richardson WS, Detsky AS: Users guides to the medicalliterature. VII. How to use a clinical decision analysis. A. Arethe results of the study valid? Evidence-Based MedicineWorking Group. JAMA 273:1292-1295, 1995

52. Rowbotham M, Harden N, Stacey B, Bernstein P, Magnus-Miller L: Gabapentin for the treatment of postherpetic neural-gia: a randomized controlled trial. JAMA 280:1837-1842, 1998

53. Rull JA, Quibrera R, Gonzalez-Millan H, Lozano CO:Symptomatic treatment of peripheral diabetic neuropathywith carbamazepine (Tegretol): double blind crossover trial.

Diabetologia 5:215-218, 1969

54. Serpell MG: Gabapentin in neuropathic pain syndromes:a randomised, double-blind, placebo-controlled trial. Pain99:557-566, 2002

55. Sindrup SH, Jensen TS: Efficacy of pharmacologicaltreatments of neuropathic pain: an update and effect re-lated to mechanism of drug action. Pain 83:389-400, 1999

56. Sindrup SH, Madsen C: The effect of tramadol in painfulpolyneuropathy in relation to serum drug and metabolitelevels. Clin Pharmacol Ther 66:636-641, 1999

57. Spiegel BM, Targownik L, Dulai GS, Gralnek IM: The

cost-effectiveness of cyclooxygenase-2 selective inhibitors inthe management of chronic arthritis. Ann Intern Med138:795-806, 2003

58. Tsevat J, Goldman L, Soukup JR, Lamas GA, Connors KF,Chapin CC, Lee TH: Stability of time-tradeoff utilities in sur-vivors of myocardial infarction. Med Decis Making 13:161-165, 1993

59. Turkington RW: Depression masquerading as diabeticneuropathy. JAMA 243:1147-1150, 1980

60. Volmink J, Lancaster T, Gray S, Silagy C: Treatments forpostherpetic neuralgiaa systematic review of randomizedcontrolled trials. Family Practice 13:84-91, 1996

61. Vrethem M, Boivie J, Arnqvist H, Holmgren H, LindstromT, Thorell LH: A comparison of amitriptyline andmaprotilinein the treatment of painful polyneuropathy in diabetics andnondiabetics. Clin J Pain 13:313-323, 1997

62. Watson CP, Chipman M, Reed K, Evans RJ, Birkett N:Amitriptyline versus maprotiline in postherpetic neuralgia:a randomized, double-blind, crossover trial. Pain 48:29-36,1992

63. Watson CP, Evans RJ, Reed K, Merskey H, Goldsmith L,Warsh J: Amitriptyline versus placebo in postherpetic neu-ralgia. Neurology 32:671-673, 1982

64. Watson CP, Vernich L, Chipman M, Reed K: Nortriptylineversus amitriptyline in postherpetic neuralgia: a random-ized trial. Neurology 51:1166-1171, 1998

65. Wiffen P, Collins S, McQuay H, Carroll D, Jadad A, MooreA: Anticonvulssants drugs for acute and chronic pain. Co-chrane Database Syst Rev 3, 2005, CD001133

66. Wong IC, Lhatoo SD: Adverse reactions to new anticon-vulsant drugs. Drug Saf 23:35-56, 2000


dor neuripática

Documents