REVIEW ARTICLE

Drugs 2000 Sep; 60 (3): 555-574 0012-6667/00/0009-0555/$25.00/0 Adis International Limited. All rights reserved.

Oral versus Topical NSAIDs in Rheumatic Diseases

A Comparison
Catherine A. Heyneman,1 Cara Lawless-Liday1 and Geoffrey C. Wall2
1 Department of Pharmacy Practice and Administrative Sciences, Idaho State University College of Pharmacy, Pocatello, Idaho, USA 2 Department of Pharmacy Practice, Drake University College of Pharmacy, Des Moines, Iowa, USA

Contents
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Rationale for the Use of Topical Nonsteroidal Anti-Inflammatory Drugs (NSAIDs) 1.2 Overview of Rheumatic Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Mechanism of Action of NSAIDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Objective Criteria for Topical Administration . . . . . . . . . . . . . . . . . . . . . 2. Absorption and Distribution of Topically Administered NSAIDs . . . . . . . . . . . . . . 2.1 Basic Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 In Vivo Animal Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Ex Vivo Studies on Human Skin Tissue . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Human In Vivo Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Penetration Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Efficacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Topical NSAIDs versus Placebo and Active Controls . . . . . . . . . . . . . . . . . 3.2 Topical NSAIDs versus Oral NSAIDs . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Cost Effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555 556 556 559 559 559 559 559 560 560 560 562 564 564 568 570 570 571

Abstract

Nonsteroidal anti-inflammatory drugs (NSAIDs) are among the most commonly prescribed drugs worldwide and are responsible for approximately onequarter of all adverse drug reaction reports. NSAIDs are widely prescribed for patients with rheumatic disease a population at increased risk for serious gastrointestinal (GI) complications. Topical administration of NSAIDs offers the advantage of local, enhanced drug delivery to affected tissues with a reduced incidence of systemic adverse effects, such as peptic ulcer disease and GI haemorrhage. NSAIDs administered topically penetrate slowly and in small quantities into the systemic circulation; bioavailability and maximal plasma NSAID concentration after topical application are generally less than 5 and 15%, respectively, compared with equivalent oral administration. Product formulation may have a dramatic impact, not only on absorption rates but also on penetration depth.

556

Heyneman et al.

Compared with oral administration, topical application leads to relatively high NSAID concentrations in the dermis. Concentrations achieved in the muscle tissue below the site of application are variable, but are at least equivalent to that obtained with oral administration. NSAIDs applied topically do reach the synovial fluid, but the extent and mechanism (topical penetration versus distribution via the systemic circulation) remain to be determined. In addition, marked interindividual variability was noted in all studies; percutaneous absorption may be strongly influenced by individual skin properties. In general, interpretation of clinical studies measuring efficacy of topical NSAIDs in rheumatic disease states is difficult because of a remarkably high placebo response rate, use of rescue paracetamol (acetaminophen), and significant variability in percutaneous absorption and response rates between patients. Overall efficacy rates attributable to topical NSAIDs in patients with rheumatic disorders ranged from 18 to 92% of treated patients. Topically applied NSAIDs have a superior safety profile to oral formulations. Adverse effects secondary to topical NSAID application occur in approximately 10 to 15% of patients and are primarily cutaneous in nature (rash and pruritus at site of application). GI adverse drug reactions are rare with topically applied NSAIDs, compared with a 15% incidence reported for oral NSAIDs. Available clinical studies suggest, but do not document, equivalent efficacy of topical over oral NSAIDs in rheumatic diseases.

1. Background
1.1 Rationale for the Use of Topical Nonsteroidal Anti-Inflammatory Drugs (NSAIDs)

The concept of transdermal application of nonsteroidal anti-inflammatory drugs (NSAIDs) is not new. Phenylbutazone cream was utilised over 30 years ago to treat superficial thrombophlebitis.[1] The renewed interest in topical delivery of NSAIDs for patients with rheumatic diseases stems from the potential advantages of transdermal administration. In theory, an NSAID applied topically could achieve therapeutic concentrations in the tissues subjacent to the site of application while maintaining low serum concentrations. This could provide numerous potential benefits to patients, including avoidance of gastrointestinal (GI) tract and firstpass metabolism, as well as reduced risk of serious adverse events related to elevated serum NSAID concentrations. Avoidance of the GI tract should mitigate the common direct toxicities of nausea, vomiting, dyspepsia and diarrhoea which occur secondary to high local concentrations of NSAID in the alimentary tract.
Adis International Limited. All rights reserved.

NSAID-mediated toxicity is often dose related. Thus, reduction in serum concentrations should also lessen the risk of potentially serious systemic adverse effects secondary to NSAID-induced prostaglandin inhibition: acute renal insufficiency, nephrotic syndrome, NSAID gastropathy, prolonged bleeding time, and fluid retention. Finally, topical application should mitigate the risk of drug-drug interactions, such as NSAID-mediated protein binding displacement of warfarin. In 1980, benzydamine was the first topical NSAID to be licensed in the UK.[2] Since that time, several other topical NSAIDs have been licensed in Europe, Japan and South Africa for the treatment of rheumatic diseases, and some of these transcutaneous formulations are currently available without a prescription.[1,2] A list of commercially-available topical NSAID preparations is presented in table I. The use of topical NSAIDs for acute musculoskeletal conditions has been reviewed extensively;[3,4] this review will therefore focus exclusively on the efficacy and safety of topical NSAIDs in chronic rheumatic diseases, including osteoarthritis, bursitis, tendinitis and epicondylitis. Unfortunately, no clinical trials have assessed the effectiveDrugs 2000 Sep; 60 (3)

Oral versus Topical NSAIDs in Rheumatic Diseases

557

Table I. Commercially available preparations of topical nonsteroidal anti-inflammatory drugsa Generic name Benzydamine Bufexamac Trade name Difflam Tantum Croxaryl Droxaryl Malipuran Parfenac Parfenac-Fettsalbe Dealgic Diclofenac GNR Gel Dicloreum Flector Flogofenac Solarase Voltaren Emulgel Voltarol Emulgel Dicloreum Tissugel Flector EP Tissugel Flector EP Gel Etofen Gel 5% Etofen Gel 10% Reumagel Flogoprofen Traxam Napageln Dolinac Gel TransAct LAT Froben Aciril Arfen Artrene Bufen Brufort Dolgit Dolofast Ibudros Ibugel Ibuleve Iprogel Optifen Gel Optifen Gel forte Proflex Adco-Indomethacin Gel Elmetacin Indocid Actron Algogel Atrosilene Gel Fastum Flexen Ketalgin Ketartrium Ketoprofene GNR Strength 3% 5% 5% 5% 5% 5% 5% 1% 1% 1% 1% 1% 3% 1% 1% 180mg 185.5mg 1% 5% 10% 5% 5% 3% 1% 3% 40mg 0.25% 10% 10% 20% 5%, 10% 10% 5% 10% 5% 5% 5% 5% 5% 10% 5% 1% 1% 1% 2.5% 1% 5% 1%, 2.5% 2.5%, 5% 2.5% 2.5% 5% Dosage form Gel Cream Cream Cream Cream Cream Ointment Gel Gel Gel Gel Gel Gel Emulsion gel Emulsion gel Plaster Plaster Gel Gel Gel Gel Gel Gel Ointment Gel Patch Spray Gel Gel Gel Cream Ointment Cream Gel Ointment Gel Gel Gel Gel Gel Cream Gel Spray Gel Gel Gel Gel Gel Gel Cream Gel Ointment

Diclofenac

Diclofenac hydroxyethyl-pyrrolidine (DHEP) Etofenamate

Felbinac

Flurbiprofen Ibuprofen

Indomethacin

Ketoprofen

Adis International Limited. All rights reserved.

Drugs 2000 Sep; 60 (3)

558

Heyneman et al.

Table I. Contd Generic name Ketoprofen Trade name Ketoprofene Ibi KLS Gel-Spray Oki Orucote Orudis Orudis Profenid Salient Thermalgen Aprenin Artroxen Gel Napreben Naprossene Naprosyn Gel Neoeblimon Numidan Gel Prexan Ticoflex Gel Xenar Xenar Niflugel R Tanderil Antiflog Atroxicam Brexivel Bruxicam Ciclafast Clevian Gel Euroxi Feldene Lipo Feldene Feldene Topico Gel Flodol Flogobene Lampoflex Oxicam Piroxicam GNR Pivaloxicam Reucam Reudene Reumagil Riacin Roxene Roxenil Roxedin Unicam Mobilat Gel Strength 1% 15% 15% 2.5% 2.5%, 5% 1% 2.5% 1% 2.5% 10% 10% 10% 10% 10% 10% 12% 10% 10% 10% 5% 2.5% 10% 1% 1% 1% 1% 2% 1% 1% 0.5% 1% 0.5% 1% 1% 1% 1% 1% 2% 1% 1% 1% 1% 1% 1% 1% 2% 2% salicylic acid, 0.2% glycosaminoglycan, 1% suprarenal extract 1% 5% Dosage form Cream Gel-spray Gel Gel Gel Cream Gel Cream Cream Gel Gel Gel Gel Gel Gel Gel Ointment Gel Gel Ointment Gel Ointment Gel Cream Cream Cream Cream Gel Gel Cream Cream Gel Gel Cream Cream Gel Cream Cream Cream Gel Cream Cream Cream Cream Gel Cream Gel

Naproxen

Niflumic acid Oxyphenbutazone Piroxicam

Salicylic acid/ glycosaminoglycan/ corticosteroid combination Suprofen Sulprotin Tolmetin Tolectin Gel a All percentages expressed are percentage weight to weight.

Ointment Gel

Adis International Limited. All rights reserved.

Drugs 2000 Sep; 60 (3)

Oral versus Topical NSAIDs in Rheumatic Diseases

559

ness of topically applied NSAIDs in rheumatoid arthritis, even though it has a prevalence rate of 1% in the adult population and oral NSAIDs are a mainstay of therapy.[5]
1.2 Overview of Rheumatic Diseases

various NSAIDs in vitro tends to be proportional to their anti-inflammatory potency in vivo.[15] The mode of COX inhibition by NSAIDs is complex and varies considerably between agents in this class.[16]
1.4 Objective Criteria for Topical Administration

The appellation rheumatic diseaseis an umbrella term encompassing a remarkably broad range of disorders characterised by varying degrees of inflammation, tissue damage and loss of function.[6] The inflammatory process is a normal response to tissue insult. However, in chronic disease states such as rheumatic disorders inflammation itself can fuel progressive disease manifestations.[7] This review will focus primarily on investigations in patients with osteoarthritis (OA) and regional rheumatic pain syndromes such as bursitis, tendinitis, tenosynovitis and epicondylitis. OA is one of the most common rheumatic disorders. This chronic, painful disease state is estimated to affect 15.8 million adults in the United States alone,[8] and is the principal source of pain and disability in the elderly.[9] OA is characterised by degeneration and loss of articular cartilage, manifesting as joint pain, stiffness and limitation of movement.[8] Regional rheumatic pain syndromes are classified as painful, inflammatory processes occurring in muscles, tendons, joints, cartilage, ligaments, fascia, bone and nerve tissues.[10] First-line treatment of localised rheumatic pain and joint stiffness generally consists of oral paracetamol (acetaminophen) and/or topical counterirritants, such as capsaicin.[7,8,10,11] NSAIDs are effective,[8] but are generally considered second-line therapy because of their adverse effect profile.[12,13] As mentioned earlier, adverse effects secondary to NSAID use include GI tract, renal, cutaneous, CNS, hepatic and haematopoietic system toxicities.[7]
1.3 Mechanism of Action of NSAIDs

In order to confirm the validity of the postulate that topical NSAID administration is superior to the oral route, the following basic criteria should be met:[2] therapeutic concentrations of NSAID must be achieved in the target tissues efficacy must be demonstrated in well-designed clinical trials the risk for serious adverse events secondary to topical NSAID therapy must be less than that of oral treatment, topical administration of NSAIDs must be cost effective. This review will attempt to address all of these criteria as they apply to the treatment of rheumatic diseases. 2. Absorption and Distribution of Topically Administered NSAIDs
2.1 Basic Principles

The primary mechanism of action of NSAIDs is reversible inhibition of the cyclo-oxygenase (COX) enzyme responsible for synthesis of prostaglandins, which are mediators of the inflammatory process.[14] The relative potency of COX inhibition by
Adis International Limited. All rights reserved.

The skin acts as an efficient barrier to the passage of materials into and out of the body.[17] The keratinised stratum corneum is the outer, horny layer which consists primarily of multilaminate hydrophobic and hydrophilic channels composed of nonpolar fatty acid chains and polar head groups.[18,19] Thus, the ideal drug candidate for transdermal delivery would have a low molecular weight, be highly potent, and have both hydrophobic and hydrophilic properties.[17] In addition, the best delivery system would release the drug to the skin at a rate lower than the maximum rate of skin transport. This would control for variability in skin permeability between individuals and ensure a constant release rate.[17,20] In several in vitro studies, a linear relationship has been established between the initial topical NSAID concentration and the release rate from the vehicle,
Drugs 2000 Sep; 60 (3)

560

Heyneman et al.

suggesting a diffusion-controlled model.[20,21] Vehicle pH and penetration enhancers such as polyethylene glycol, limonene, emulsion preparations or iontophoresis can also have dramatic effects on NSAID release rates, increasing penetration rates by up to 75-fold.[21-26]
2.2 In Vivo Animal Studies

In order to treat patients rationally with topically applied NSAIDs, the issue of local enhanced topical delivery of NSAIDs must be addressed. This concept postulates that local accumulation of drug in target tissues occurs by direct diffusion to a greater degree than could have resulted by prior absorption and redistribution through the cutaneous vasculature.[2] This issue has been addressed directly by McNeill et al.,[27] who administered equivalent 50g doses of titrated piroxicam either intravenously or topically to the right shoulder of male rats. In the muscle tissue subjacent to the topical piroxicam application, 2 distinct piroxicam peaks were noted at 4 and 12 hours after application. The contralateral muscle tissue from the nondosed shoulder produced a single peak at 12 hours. The authors interpretation of this data was that the first peak could only be accounted for by topical piroxicam delivery; the second peak was attributed to redistribution from the general blood supply. The concentrations of piroxicam achieved in the acromiotrapezius muscle subjacent to the treated shoulder at 2 hours after topical application exceeded that found in the same muscle of rats given equivalent intravenous piroxicam by 14-fold. This work was confirmed by Mikulak et al.[28] These authors reported that topical application of indomethacin produced maximal tissue concentrations in the subjacent deltoid muscle, joint capsule and rotator cuff 2 to 10 times greater than that achieved with equivalent oral dosing. Singh and Roberts[29] investigated the relative epidermal permeabilities of salicylate, indomethacin, naproxen, diclofenac and piroxicam solutions in anaesthetised rat dermis. These authors concluded that local direct penetration was evident for all NSAIDs up to a depth of about 3 to 4mm below
Adis International Limited. All rights reserved.

the applied site, with distribution to deeper tissues being mainly though the systemic blood supply. While this information seems to contradict the studies by Mikulak et al.[28] and McNeill et al.[27] with respect to deeper tissue penetration, it is important to note that the drugs utilised in the study designed by Singh and Roberts[29] were in aqueous solution, while Mikulak et al.[28] utilised a lipophilic penetration enhancer and McNeill et al.[27] used a carbopol gel base. Thus, product formulation may have a dramatic impact, not only upon absorption rates, but also on penetration depth.
2.3 Ex Vivo Studies on Human Skin Tissue

Ex vivo studies utilising cadaver skin tissue to investigate transdermal penetration of NSAIDs have determined that occlusion of the skin surface can enhance drug penetration by up to 2-fold.[30] In a study of 7 different NSAIDs, Cordero et al.[31] determined that the flux across human cadaver skin was related primarily to the lipophilic character of the drug. Roy and Manoukian[32] demonstrated that the flux of ketorolac through skin tissue in vitro decreased exponentially as the solution pH was raised form 3.5 to 7.0. This is consistent with the findings of Cordero et al.,[31] given that this pH shift would increase the unionised percentage of ketorolac by several orders of magnitude, making it more lipophilic. In essence, ex vivo studies suggest that occlusion can enhance NSAID skin penetration, lipophilicity is an important characteristic, and the pH of the vehicle is extremely important. However, these data must be interpreted with caution, as cadaver skin obviously has no functional blood supply and is therefore an incomplete model for true skin penetration.
2.4 Human In Vivo Studies

Many investigators have attempted to quantitate maximum plasma concentrations (Cmax) and the time required to reach this blood level (tmax) after transdermal administration of NSAIDs in healthy humans.[32-40] These reports have been summarised in table II for both single dose and multiple dose
Drugs 2000 Sep; 60 (3)

Oral versus Topical NSAIDs in Rheumatic Diseases

Adis International Limited. All rights reserved. Drugs 2000 Sep; 60 (3)

Table II. Summary of single and multiple dose topical nonsteroidal anti-inflammatory drug absorption studies in healthy individuals. When the trial also involved an oral or intramuscular agent, only the results relating to the topical agent(s) are shown Drug Dosage Route n Design Mean plasma Cmax (g/ml) SD or SE or %CV Ratio of Cmax observed to expected Cmax from equivalent oral administrationa 0.007 0.003 0.080 0.016 0.022 0.081 0.032 0.005 0.001 0.001 0.011 Not available 0.044 Mean plasma tmax (hours) SD or SE or %CV Reference

Single dose studies Diclofenac 300mg solution gel 300mg emulsion gel Ibuprofen 200mg solution gel 200mg hydrophilic ointment 200mg oil in water emulsion 174mg in vehicle A 239mg in vehicle B 205mg in vehicle C 1.5g in device A (20cm2) 1.5g in device B (20cm2) 0.86g in device C (7.5cm2) 50mg gel 15mg gel

Topical over 400cm2, occluded Topical over 360cm2 occluded

12 6 6 6

Randomised, 2way, crossover Randomised, 3way, crossover

0.081 0.043 (SD) 0.039 0.017 (SD) 1.40 0.40 (SD) 0.28 0.11 (SD) 0.39 0.19 (SD)

3.50 0.52 (SD) 7.17 1.59 (SD) 4.83 1.83 (SD) 7.0 1.67 (SD) 6.8 2.2 (SD) 5.3 38% (CV) 7.1 47% (CV) 22.2 + 10% (CV) 8.4 1.3 (SD) 12.4 1.3 (SD) 8.8 1.7 (SD) Below detection 23 1.7 (SE)

33 34

Ketorolac

Ketorolac

9 9 9 Topical over variable 10 surface area according to device, 10 occlusion unclear 10 Topical over 70cm2 (occlusion not noted) Topical over 289cm2, not occluded Topical over 200cm2, occluded Topical over 150cm2, occluded Topical (surface area not specified), occlusion not noted Topical over 100cm2 on back, arm or knee, not occluded 5 8

Topical over 20cm2, occluded

Randomised, 3way, crossover Randomised, 3way, crossover

1.265 45% (CV) 0.696 72% (CV) 0.092 63% (CV) 0.20 0.15 (SD) 0.18 0.06 (SD) 0.82 0.24 (SD)

35

32

Ketoprofen Piroxicam

Open Open

Below detection 0.147 0.034 (SE)

36 37

Multiple dose studies Diclofenac 80mg bid for 7.5 days Diclofenac hydroxyethylpyrrolidine (DHEP) Ketoprofen 185.5mg plaster bid for 7 days 50mg gel bid for 7 days 30mg gel q 6 hours 25 doses

12 10

Open Randomised, 2way, crossover

0.0189 0.0172 (SD) 0.017 0.013 (SD) 0.028 0.013 (SD)

0.006 0.002 0.019

2.2 2.3 (SD) 5.4 3.7 (SD) 3.1 0.7 (SD)

38 39

0.119 + 78% (CV) from 0.066 3.00 + 73% (CV) back from back 2.67 + 75% (CV) 0.119 + 58% (CV) from 0.066 arm from arm 0.085 + 65% (CV) from 0.047 3.04 + 53% (CV) knee from knee a Estimated from manufacturers reported Cmax values after oral administration, corrected for dosage differences. Cmax = maximum plasma concentration; SE = standard error; SD = standard deviation; tmax = time to Cmax; %CV = percentage coefficient of variance. 24 Randomised, 3way, crossover

40

561

562

Heyneman et al.

studies, along with a calculation comparing the reported mean Cmax values to expected plasma concentrations after equivalent oral administration. These data indicate that systemic absorption after topical administration produces peak plasma values less than 10% of that obtained after oral administration (range 0.2 to 8.0%). Another important point gleaned from table II is the significant interand intra-individual variability, which is consistent with that seen in the ex vivo data.[31] This variability has been attributed to individual skin differences between study participants, including hydration status,[41] permeability characteristics[31,38] and individual differences in subcutaneous vasculature.[41] Shah et al.[40] compared serum ketoprofen concentrations after equivalent gel applications to the forearm, back or knee in a crossover design (see table II). These authors found that anatomical site of application had little effect on systemic absorption. The study by Seth[34] comparing equivalent doses of ibuprofen applied to equivalent skin surface areas emphasises the importance of the vehicle on percutaneous drug absorption. This author demonstrated a 5-fold enhancement of maximal serum concentration by changing the drug formulation from an ointment base to a gel base. Overall, given that different NSAIDs, vehicles, methods of occlusion and/or permeability enhancers were utilised in every study described in table II, it is difficult to draw any conclusions regarding superior permeability of any one NSAID formulation. As demonstrated by the available in vivo studies (see table II), topical NSAIDs penetrate slowly and in small quantities into the general circulation. Marked interindividual variability was noted in all studies. Reported maximal plasma concentrations of NSAIDs applied topically are generally less than 15% of the concentration which would be expected from equivalent oral administration. The reported maximal serum concentrations after topical administration are uniformly below the accepted therapeutic drug concentrations for NSAIDs, which is advantageous when viewed from the vantage point of systemic toxicity. The time required to achieve Cmax is approximately 10 times longer than that of
Adis International Limited. All rights reserved.

equivalent oral administration, ranging from 2.2 to 23 hours. Bioavailability studies suggest that NSAIDs administered topically achieve only 3 to 5% of the total systemic absorption when compared with oral administration.[33,42,43] Anatomical site of application does not appear to influence plasma concentrations,[40] but the formulation can enhance systemic absorption by as much as 5-fold.[34] Steady state is achieved fairly rapidly after tmax (generally within 2 to 5 days of initiation of repeated topical application),[41,42] and plasma concentrations after repeated administration have been reported to rise to 2.5 times the Cmax after a single application.[42]
2.5 Penetration Studies

The concept of local enhanced topical delivery of NSAIDs requires proof not only that the drug reaches the systemic circulation, but that it reaches higher and therapeutically effective concentrations in the tissues localised below the site of application. Verification of local enhanced topical delivery is essential, given the proposed rationale behind the use of topical NSAIDs that enhanced, localised concentrations minimise systemic toxicity while maintaining efficacy. Muller et al.[38] attempted to address the issue of depth of diclofenac penetration in vivo. In this study, 12 healthy volunteers were given topical diclofenac foam (diclofenac 80mg over 200cm2 applied to the shaved surface of the thigh with occlusion) twice daily for 8 days. Immediately prior to the last topical dose, 2 microdialysis probes were inserted to a tissue depth of 11.7 0.4mm without local anaesthesia into the medial vastus muscle at least 4cm lateral from the margin of the application site. Concentration versus time profiles were obtained for plasma and interstitial fluid of skeletal muscle. Significant interindividual variability was noted. The mean Cmax in plasma was 18.75 4.97 g/L, and corresponding interstitial concentrations in skeletal muscle were 219.68 66.36 g/L (12fold higher than the plasma level). A similar study was conducted by Tegeder et al.[44] These investigators administered ibuprofen 800mg either orally or topically over 323cm2 to the
Drugs 2000 Sep; 60 (3)

Oral versus Topical NSAIDs in Rheumatic Diseases

563

thighs of 11 healthy volunteers in a 2-way crossover design, then monitored serum and tissue concentrations by microdialysis probes inserted into the thigh to a depth of 4 to 5mm (dermis) and 25 to 30mm (muscle). Topical administration led to 22.5-fold greater concentrations of ibuprofen in the dermis than those obtained after oral administration (p value not reported). Ibuprofen concentrations achieved in the muscle after topical administration were almost identical to those obtained after oral administration. There was significant interindividual variability (for example, muscle ibuprofen concentrations after topical administration exceeded those from oral administration by over 2 orders of magnitude in 1 volunteer). In approximately 50% of the volunteers, ibuprofen concentrations in the muscle tissue after topical administration were increased over concentrations after oral administration. The authors attributed this variability to differences in the microvasculature at the site of drug administration. Keeping in mind that the interstitial drug concentration in the muscle tissue measured by Muller et al.[38] was 11mm deep and 4cm lateral to the application site, these results directly contradict the in vivo animal study by Singh and Roberts,[29] which postulated that distribution to tissues deeper than 3 to 4mm below the applied site were accounted for by the systemic blood supply. The authors note that the microdialysis method employed measured only the pharmacologically active unbound drug fraction; total tissue drug concentrations were probably higher. The study by Tegeder et al.[44] suggests that, for approximately 50% of patients, local enhanced topical delivery of NSAIDs accounts for tissue concentrations of drug that are above that which can be obtained by equivalent oral administration up to 25 to 30mm subjacent to the site of application. In an elegant, nonblind study design, Dominkus et al.[45] compared topical administration of ibuprofen 375mg in a gel formulation applied 3 times daily for 3 days to oral administration of ibuprofen 600mg given twice daily for 3 days in 17 patients with degenerative knee disorders. Ibuprofen ad Adis International Limited. All rights reserved.

ministration occurred in the time period immediately preceding knee surgery. Twelve patients received the topical preparation (1125 mg/day for 3 days) and 5 received the oral formulation (1200 mg/day for 3 days). Samples of blood, synovial fluid, muscle, fasciae and subcutaneous tissue were obtained during the operation, 15 hours after the last administration of ibuprofen. Mean plasma concentrations of ibuprofen were 1.0 0.5 mg/L in patients receiving the topical preparation, and 1.6 1.3 mg/L in patients receiving oral ibuprofen (p = 0.17). Given that the half-life of ibuprofen is 1.78 to 2.5 hours,[46] at least 6 half-lives elapsed between the last topical or oral dose and the time the serum and tissue samples were removed for analysis, making it very difficult, if not impossible, to compare the blood concentrations. However, it is interesting to note that similar ibuprofen concentrations were achieved in the 2 treatment groups in the synovial fluid, fasciae and muscle tissues. The subcutis was the only tissue in which the ibuprofen concentration derived from topical application exceeded that obtained after oral administration (5.4 3.3 vs 1.3 0.3 g/g; p = 0.03). In a recent randomised, parallel investigation, Rolf et al.[47] compared ketoprofen concentrations in plasma, synovial fluid and intra-articular tissues in 100 patients undergoing knee arthroscopy. Forty patients received a single topical plaster containing ketoprofen 30mg, 30 received multiple 30mg plaster applications over 5 days and 30 received ketoprofen 50mg orally. Ketoprofen concentrations were measured in plasma, synovial fluid, synovial tissue, meniscus and cartilage at time points up to 14 hours after drug administration. Interestingly, while synovial fluid ketoprofen concentrations averaged 70 to 80% of the plasma concentrations after topical application, the meniscus and cartilage concentrations were elevated 20- to 30-fold over those in plasma in the same patients. Median ketoprofen concentrations achieved in meniscus and cartilage tissues after topical administration (single or multiple application) were 4.1- to 6.8-fold higher than those achieved after oral administration.
Drugs 2000 Sep; 60 (3)

564

Heyneman et al.

Dominkus et al.[45] reported that tissue and synovial fluid ibuprofen concentrations achieved after topical application exceeded the in vitro IC50 (concentration of ibuprofen required to inhibit 50% of enzyme activity) for prostaglandin synthetase[48] even 6 half-lives after last administration of drug. IC50 values for ibuprofen inhibition of COX-1 and COX-2 have been determined in vitro to be 0.8 to 3.8 mg/L and 0.0006 to 0.0012 mg/L.[16] COX-2 is the isoform primarily associated with inflammation;[16] thus, ibuprofen applied topically appears to exceed the threshold quantity required for anti-inflammatory activity. However, Gierse et al.[16] have recommended caution when interpreting the COX IC50 data, because of the complex and distinct mechanisms of enzyme inhibition of each COX isoform by NSAIDs (including competitive, covalent and time-dependent variable binding mechanisms). These authors state that comparison of inhibitory activity on COX-1 and COX-2 using IC50 ratios [are of] questionable validity.[16] Judicious interpretation of IC50 data is reasonable, then, particularly when considering that COX-2 is an inducible enzyme which increases in concentration during the inflammatory response.[8] As noted with ibuprofen,[48] topically-applied diclofenac hydroxyethylpyrrolidine (DHEP) has also been detected in the synovial fluid of patients with monolateral knee joint effusion by Gallacchi and Marcolongo.[49] These authors treated 8 patients with OA of the knee with plasters containing 180mg DHEP twice daily for 4.5 days. Four hours after the last application, synovial fluid from the treated knee joint was collected. Plasma diclofenac concentrations were also assessed at 1, 4 and 8 hours after the last topical application. The mean concentration of diclofenac detected in the synovial fluid was 30% of that found in the plasma at the same 4-hour time point (1.02 0.38 vs 3.62 1.05 g/L; p < 0.05). It has been postulated that the major site of action for NSAIDs in the treatment of arthralgias is probably within the synovial compartment.[50] These studies prove that topically applied NSAIDs can indeed reach the synovial fluid. Orally administered NSAIDs generally result in synovial drug concen Adis International Limited. All rights reserved.

trations approximately 60 to 80% of the mean plasma concentration;[47,50] this difference in concentration has been attributed to lower albumin levels in synovial fluid compared with plasma.[15] It is interesting to note that in the study conducted by Dominkus et al.,[45] the mean NSAID concentration in the synovial fluid actually exceeded that found in the plasma, although statistical tests were not applied to this data. Given the relatively slow transfer rate of NSAIDs into and out of the synovial compartment,[15] the sustained synovial NSAID levels observed,[49] and the fact that the levels in the study conducted by Dominkus et al.[45] were taken 15 hours after the last topical administration, it is not possible to state conclusively that topically applied NSAIDs are capable of concentrating in the synovium. However, Rolf et al.[47] provide persuasive evidence that ketoprofen applied topically can concentrate in the intra-articular tissues, and that less vascularised tissues (cartilage and meniscus) may actually act as a drug reservoir. In summary, local enhanced topical delivery of NSAIDs does occur. However, the maximal depth to which it occurs varies among individuals. Topically applied NSAIDs do reach the synovial fluid compartment and can concentrate in intra-articular tissues. However, without direct, objective measurement of contralateral tissue and synovial fluid drug concentrations in well-designed trials, a determination of the mechanism of NSAID transport whether it is direct or via the systemic circulation or both cannot yet be made. While essentially all tissue concentrations of topically applied NSAIDs result in subjacent tissue concentrations which exceed the IC50 for COX-2, this data must be interpreted cautiously based upon variable and complex mechanisms of COX enzyme inhibition.[16] 3. Efficacy
3.1 Topical NSAIDs versus Placebo and Active Controls

Table III includes a summary of published clinical trials which compared topically applied NSAIDs with vehicle controls or to other topical NSAIDS
Drugs 2000 Sep; 60 (3)

Table III. Summary of clinical trials investigating the efficacy of topical nonsteroidal anti-inflammatory drugs (NSAIDs) in rheumatic conditions: versus placebo and active controls Active drug therapy Formulation Treatment regimen Length Total Inclusion of number of diagnoses study participants (days) (number completing trial) Trial Efficacy vs control at design end of trial Overall clinical status % Adverse reactions (% requiring dropout) Rescue med (daily maximum allowed) Ref.

Oral versus Topical NSAIDs in Rheumatic Diseases

Adis International Limited. All rights reserved. Drugs 2000 Sep; 60 (3)

pain ratings functional patient ratings ratings Comparative clinical trials: topical NSAID vs vehicle control Flurbiprofen 40mg LAT 40mg patch over 14 80 (70) patch 136cm2 bid over affected shoulder, occluded Flurbiprofen 40mg LAT 40mg patch over 14 104 (89) patch 136cm2 bid over affected area occluded Scapulohumeral periarthritis with mild-moderate pain Local, nonarticular form of rheumatism with moderate-severe pain (epicondylitis, tendonitis, bursitis) 12 epicondylitis, 30 scapulohumeral periarthritis, 10 gonalgia, 8 metatarsalgia Osteoarthritis of the hip and/or knee r, db, No p difference

physician active control ratings drug 8.8 (0) Paracetamol 51 [acetaminophen] (up to 2 g/day) Paracetamol 52 (up to 4 g/day)

No No No 22.2 difference difference difference (0)

r, db, No p difference

No No Flurbipro- 15 difference difference fen; p = (3.8) 0.009

6 (3.9)

Ketorolac

0.6% aqueous solution

Coppersalicylate

DHEP

Chelated coppersalicylate gel (4.3 mg/g copper, 43.8 mg/g salicylate) 180mg plaster

30mg via iontophoresis qod (surface area not reported, site not specified), not occluded 6.5mg copper + 65.7mg salicylate to medial forearm bid (surface area not specified), not occluded 180mg over 150cm2 bid (site not specified), occluded 180mg over 150cm2 bid applied at site of inflammation, occluded

10

60 (60)

r, db, No p difference

Not Ketorolac; Ketorolac; 0 (0) assessed p < 0.02 p < 0.02

0 (0)

Not reported 53

28

116 (93)

r, db, No p difference

Not No No 85 assessed difference difference (24)

52 (1.7)

Paracetamol 54 (no maximum reported)

15

155 (142)

Osteoarthritis of r, db, DHEP; p < the knee (clinically p 0.0001 symptomatic) Local r, db, DHEP; p < inflammatory p 0.01 processes in periarticular/ tendinous and/or extra-articular sites

DHEP; p< 0.0002

DHEP; p < 0.0001

DHEP; p < 0.0001

1.3 (0)

5.2 (1.3)

DHEP

180mg plaster

14

60 (60)

Not Not DHEP; p assessed assessed < 0.01

0 (0)

0 (0)

Paracetamol 55 (no maximum reported) Paracetamol 56 (no maximum reported)

565

Table III. Cont Active drug therapy Formulation Treatment regimen Length Total Inclusion of number of diagnoses study participants (days) (number completing trial) Trial Efficacy vs control at design end of trial Overall clinical status % Adverse reactions (% requiring dropout) Rescue med (daily maximum allowed) Ref.

566

Adis International Limited. All rights reserved. Drugs 2000 Sep; 60 (3)

Diclofenac/ hyaluronate

60mg diclofenac, 14 50mg hyaluronan daily over affected joint (surface area not specified), not occluded Indomethacin 4% spray 10-20 mg/dose 3-5 14 with metered times daily over dose aerosol affected area, not unit occluded Diclofenac 2% in lecithin 50 mg/dose tid 14 organogel over affected knee, not occluded

3% diclofenac, 2.5% hyaluronan gel

119 (112)

Osteoarthritis

pain ratings functional ratings r, db, No Not p difference assessed

patient ratings Not assessed

physician ratings Not assessed

active control drug 20.3 43.3 Not allowed (0) (0)

57

30 (30)

28 peri-arthritis of shoulder, 2 epicondylitis

r, db, Indometha- Indome- Indomeco cin; p < thacin; p thacin; p 0.001 < 0.014 < 0.025 Diclofenac p = 0.05

Indomethacin; p < 0.01

6.7 (0)

0 (0)

Paracetamol 58 (up to 3 g/day) Paracetamol 59 (up to 4 g/day)

74 (70)

Osteoarthritis of r, db, Diclofenac p = 0.05 the knee (clinically p symptomatic and radiologically verified) Nonsurgical lumbago (disc syndrome, back pain, etc.) r, sb, p No difference

No No 2.7 difference difference (2.7)

0 (0)

Comparative clinical trials: topical NSAID vs topical NSAID Plaster applied bid 28 Topical indo- Plaster 30 vs 34 methacin vs (strength not on lower back (64 total) topical specified) vs (surface area not diclofenac emulsion gel specified), (strength not occluded vs 2cm gel applied qid specified) (surface area not specified), not occluded Topical 40mg LAT 40mg patch over 137 total 14a flurbiprofen patch vs 136cm2 applied (122 at vs topical 0.5% gel bid over affected end) piroxicam area vs 450mg applied over affected area (surface area not specified), not occluded

No No Not Indo- Diclodifference difference assessed meth- fenac acin group group 0 (0) 6.7 (3.3)

Paracetamol 60 (no maximum reported)

Soft tissue r, nb, LAT; rheumatism of the co p < 0.001 shoulder or elbow

LAT; p < 0.004

LAT; LAT; p < 0.001 p < 0.04

LAT group 9.0 (0.8)

Piroxicam group 6.7 (2.4)

Paracetamol 61 (no maximum reported)

Heyneman et al.

Oral versus Topical NSAIDs in Rheumatic Diseases

567

Landscape table III to be placed here.

in the treatment of rheumatic conditions.[51-63] A cursory review of the placebo-controlled trials summarised in table III would suggest that, with the possible exception of DHEP and indomethacin, topical NSAIDs are not effective in chronic inflammatory diseases. Indeed, the discouraging results reported by Shackel et al.[54] are not unexpected, given that patients with OA of the hip and/or knee were instructed to apply the copper-salicylate gel to their forearms, instead of over the painful joint. However, of the 5 studies that reported no difference in pain ratings between topical NSAID and placebo treatments,[51-54,57] 2 reported decreases in subjective pain scores by 35 to 60% from baseline which were attributable to placebo alone (p values < 0.05 compared to baseline).[51,53] This high placebo response rate is consistent with that seen in topical treatment of soft tissue injury.[3,51,64] Vaile and Davis[3] noted that the placebo response associated with topical treatment of soft tissue complaints can be as high as 60 to 80%. While the natural progression of healing may account for the high placebo rate in patients with acute musculoskeletal complaints, the high rate reported in more chronic rheumatological diseases is more puzzling. Simple rubbing of the topical preparation over the affected area may have had some positive treatment results. Two of the studies noted in table III utilised mentholated or camphorated placebo formulations, which may have contributed to the placebo effect.[51,54] The trend of significant pain resolution from baseline noted in the placebo-controlled studies continued in the comparative topical NSAID versus topical NSAID trials summarised in table III. Reported pain scores decreased from 39[62] to 70%,[63] and the percentage of patients who improved from baseline ranged from 26[61] to 87%[60] (p < 0.05 for all reported values). Looking at the comparative clinical trials summarised in table III, it is not possible to draw conclusions regarding a superior drug. This is not surprising, given that there is high interindividual variability in patient response to NSAIDs administered orally, and the lack of plaDrugs 2000 Sep; 60 (3)

Adis International Limited. All rights reserved.

Topical DHEP vs topical DDA

180mg plaster vs 1.16% emulsion gel containing 1% diclofenac

Plaster applied 14 over 150cm2 bid over affected area, occluded vs 20mg applied qid over affected area (surface area not specified), not occluded

96 vs 94 (190 total)

Localised peri-articular/ tendinous inflammations

r, nb, DHEP; p p < 0.001

Not DHEP; DHEP; DHEP DDA Not reported 62 assessed p < 0.001 p < 0.001 group group 2.1 3.2 (0) (0)

Topical BPAA vs topical diclofenac

30mg applied tid 7 20 vs 20 Osteoarthritis r, nb, No Not Not Not BPAA Diclo- Not reported 63 over affected area (40 total) p difference assessed assessed assessed group fenac (surface area not 0 (0) group specified), not 0 (0) occluded vs 40mg qid over affected area (surface area not specified), not occluded a 4 days of one drug then 4 days of the other, then 6 days of preferred treatment. bid = twice daily; BPAA = biphenyl acetic acid; co = crossover; DDA = diclofenac diethyl diammonium; db = double-blind; DHEP = diclofenac hydroxyethyl-pyrrolidine; LAT = local action transcutaneous; nb = nonblind; p = parallel; qid = 4 times daily; qod = every other day; r = randomised; sb = single-blind; tid = 3 times daily.

Gel (concentration not specified) vs emulsion gel (concentration not specified)

568

Heyneman et al.

cebo controls provided no index of internal sensitivity.[8] It is also difficult to deduce the optimal topical NSAID formulation. In the sole study which directly compared 2 different formulations of diclofenac, the total daily dosages were dramatically different (360 vs 80mg), preventing any conclusions regarding vehicle superiority from being drawn.[62]
3.2 Topical NSAIDs versus Oral NSAIDs

Objective, clinical evidence for the relative efficacy of topical versus oral NSAIDs in rheumatic disorders is scant. Methodological issues cloud the results presented here. Of the 4 studies summarised in table IV,[65-68] 1 was nonblind, 3 failed to specify the surface area of application and none utilised a crossover design. A crossover methodology, with the advantage of utilising patients as their own controls, would have helped minimise the interpatient variability that seems to plague these investigations. Further, all 4 studies chose oral NSAID dosages on the lower end of the recommended dosage range and compared these oral forms to topical preparations which have not demonstrated superiority to vehicle controls (see table IV). Documentation of compliance with the study protocol, application technique (the extent of rubbing in), occlusion and skin surface area to be covered are all important variables which need to be standardised in these studies. The methods section of the study reported by Sandelin et al.[67] stated that in bilateral cases both knees were treated with same treatment and treatment regimen. Thus, patients suffering from OA in both knees received twice the dosage (and achieved presumably twice the serum concentration) of patients with monolateral OA. However, the data analysis was not structured to control for this variable. Furthermore, exclusion criteria common to these studies included a history of peptic ulcer disease, GI haemorrhage or renal dysfunction. Thus, the safety and tolerability data may have been skewed in favour of oral NSAIDs. The issue of allowance and tracking of rescue medication is also very important. The study re Adis International Limited. All rights reserved.

ported by Dickson[68] is interesting in this respect. This author compared piroxicam gel and ibuprofen tablets in patients suffering from OA of the knee. Patient ratings of efficacy were good to excellent in 64 and 60% of patients given topical piroxicam or oral ibuprofen, respectively. However, paracetamol rescue analgesia was required in 69 and 62% of patients in the topical and oral groups, respectively. This is the only study reviewed that reported such high rescue analgesic use. Many of the efficacy studies allowed rescue paracetamol and, of the studies that tracked rescue drug use,[52,54-56,66] most reported low incidences of use (under 5%).[56,66] All studies reviewed here, while not consistently documenting superiority of topical or oral administration, reported clinical improvement attributable to topical NSAIDs ranging from 18[67] to 92%[66] of treated patients. However, the placebo effect issue looms large, since only 1 study included a topical placebo into its design.[67] In this investigation by Sandelin et al.,[67] oral diclofenac showed superior efficacy to topical placebo as measured by investigators and patients evaluation of overall efficacy (p < 0.008); the efficacy values reported for topical eltenac versus placebo, while showing a trend in favour of eltenac, did not reach statistical significance. One of the 4 studies comparing oral with topical NSAID summarised in table IV documented superiority of the topical dosage form. Martens[66] compared local action transcutaneous (LAT) flurbiprofen to oral diclofenac in the treatment of soft tissue rheumatism. In the investigators opinion, 92% of patients treated with LAT showed clinical improvement, whereas 73% of those receiving oral diclofenac showed improvement (p = 0.03). In addition, the investigators ratings of severity of pain and severity of tenderness were significantly better in the LAT group (p 0.04). Unfortunately, while this was the only comparative study to show superiority of a topical formulation, it was also the only study which was nonblind, and all assessments which reached statistical significance were those of the investigator, not of the patients.
Drugs 2000 Sep; 60 (3)

Table IV. Summary of clinical trials investigating the efficacy of topical nonsteroidal anti-inflammatory drugs (NSAIDs) in rheumatic conditions: versus oral NSAIDs Active drug therapy Formulation Treatment regimen Length Total Inclusion of number of diagnoses study participants (days) (number completing trial) Trial Efficacy vs control at design end of trial Overall clinical status % Adverse reactions (% requiring dropout) Rescue med (daily maximum allowed) Ref.

pain ratings functional patient ratings ratings Topical felbinac vs oral fenbufen Gel (concentration not specified) vs 200mg tablets (Dosage, 14 occlusion and surface area not specified) tid plus placebo tablets tid vs 200mg po tid plus placebo gel tid 40mg patch over 14 136cm2 applied over affected area bid vs 50mg po bid 275 total Osteoarthritis r, db, No of the knee p difference (clinically confirmed)

physician active ratings drug

control Oral Not reported fenbufen group 10.2 (0) 65

Not No No Topical assessed difference difference felbinac group 1.4 (0)

Topical flurbiprofen vs oral diclofenac

LAT patch containing 40mg flurbiprofen vs 50mg tablets

56 vs 53 (109 total)

Localised, soft-tissue rheumatism

r, nb, Flurbiprofen; No Not p p < 0.04 difference assessed

Flurbipro- Topical fen; p < flurbi0.03 profen group 14.0 (1.8)

Oral Paracetamol diclofenac (max not group reported) 17 (3.8) flurbiprofen group used less than diclofenac group p < 0.04 Oral Not allowed diclofenac group 24.4 (0)

66

Topical 1% gel vs eltenac vs 50mg tablets oral diclofenac

30mg applied over affected knee tid (surface area not speci fied) plus 1 placebo tablet bid vs 50mg po bid plus placebo gel applied tid 5mg applied over affected knee (surface area not specified) tid vs 400mg tid

28

126 vs 82 (208 total)

Osteoarthritis r, db, No of the knee p difference (radiographica lly confirmed)

Not No No Topical assessed difference difference eltenac group 27.0 (3.2)

67

Topical piroxicam vs oral ibuprofen

0.5% gel vs 400mg tablets

28

117 vs 118 (235 total)

Mild r, db, No difference osteoarthritis p of the knee

No No No Topical Oral difference difference difference piroxicamibuprofen group group 26.5 22.9 (5.9) (7.7)

Paracetamol (up to 4g/day) (62-69% of patients in each group required rescue analgesia)

68

bid = twice daily; db = double-blind; LAT = local action transcutaneous; nb = nonblind; p = parallel; r = randomised; tid = 3 times daily.

570

Heyneman et al.

A nonblind study by Browning and Johson,[69] in 191 elderly patients with mild to moderate OA, demonstrated that patients taking oral NSAIDs for OA could reduce their oral NSAID dose by a factor of 2 by concomitant use of piroxicam gel with no reduction in perceived efficacy. Thus, addition of a topical NSAID to an existing oral regimen may allow patients to experience equivalent pain relief while tapering down the dosage of the oral form (and thus reducing their risk of GI haemorrhage, peptic ulcer disease, etc.) In summary, the empirical evidence suggests, but does not prove, that NSAIDs administered topically are at least as efficacious as oral NSAIDs in the treatment of rheumatic diseases. Larger, placebocontrolled, double-blind studies which control for the variables discussed in detail in this section are necessary before a conclusion can be reached. 4. Safety Systemic COX inhibition by NSAIDs results in a cluster of well recognised adverse effects, particularly affecting the GI tract and kidneys.[7] Some of the adverse effects secondary to oral NSAID use are dose related.[70-72] NSAID use approximately doubles the risk of acute renal failure,[72] and a linear dose-response relationship has been established between oral NSAIDS and upper GI bleeding.[73] Thus, targeting therapeutic tissue/synovial fluid concentrations of NSAIDs via the topical route while minimising systemic exposure by lowering serum concentrations makes good sense. Figueras et al.[74] reviewed 194 adverse drug reaction reports attributed to topical NSAIDs. Of these, 95% were dermatological in nature and the remaining 5% were systemic reactions. This data is consistent with the information presented in tables III and IV. On average, adverse drug reactions (ADRs) occurred in 12% of the patients in these studies (range 0 to 85%), and approximately 75% of ADRs were cutaneous, consisting of a rash and/ or pruritus at the site of application. The mean percentage of patients reporting adverse events after topical placebos was 14.4% (range 0 to 52%). Again, these events were primarily rash and pruritus, sug Adis International Limited. All rights reserved.

gesting that the vehicle itself may be responsible for a significant proportion of the adverse cutaneous reactions. The tolerability of topical NSAIDS in the elderly has been reviewed.[72] While the majority of adverse effects are skin disorders, there have been case reports of topical NSAIDs associated with systemic effects such as GI and renal toxicities, abnormal hepatic function, hematopoietic disorders and asthma.[72] However, the incidence of these serious systemic ADRs appears to be very low; none of the studies summarised here reported a serious systemic reaction attributable to topical NSAID use. Without considerable post-marketing surveillance data, it will be difficult to assign a definite association between topical NSAID administration and systemic adverse events, since many events may occur independently, and patients may have been taking an oral NSAID at the time of their illness. A case-control study conducted by Evans et al.[75] demonstrated that once adjustments were made for the confounding effects of concomitant oral NSAID use, topical NSAID administration was not significantly associated with upper GI bleeding and perforation. In summary, topically applied NSAIDs are safer than orally administered NSAIDs. ADRs can be expected in approximately 10% of patients, the vast majority being localised pruritus and/or rash at the site of administration that resolves quickly upon discontinuation of the product. Cross-sensitivity between different topical NSAIDs has been established.[76] Skin reactions secondary to oral NSAID use have been reported in 5 to 10% of patients.[77] GI ADRs after topical NSAID use are rare; in comparison, the incidence of serious GI events associated with oral NSAIDs is 15%.[78] 5. Cost Effectiveness By the year 2030, people over the age of 65 are expected to constitute 17% of the US population and to account for approximately 40% of total drug expenditures.[79] OA is the most common of the rheumatic diseases, and the overall disease prevalence increases with age: 10 to 20% of people over
Drugs 2000 Sep; 60 (3)

Oral versus Topical NSAIDs in Rheumatic Diseases

571

65 have clinical OA of the knees and hips, and over half of these have radiographic evidence of OA.[8,80] Currently, over 50% of oral NSAID prescriptions are written for OA.[15] Furthermore, increasing age is associated with an increased likelihood of adverse events after oral NSAID therapy, particularly with respect to peptic ulcer disease. The overall odds ratio for the risk of serious GI toxicity associated with oral NSAID administration is 2.74 [95% CI (confidence interval): 2.54 to 2.97]; the risk assessed in people aged 60 years or greater (compared with those under 60 years taking NSAIDs) has been estimated to be 5.5 (95% CI: 4.6 to 6.6).[81] Thus, the issue of cost effectiveness of oral versus topical administration of NSAIDs is an important consideration. Topical NSAIDs have been considered a costly alternative to oral formulations.[64,82] However, in the US alone, oral NSAID-induced GI damage has been reported to account for 42% of hospital admissions of patients with rheumatic diseases and for one-third of the total cost of rheumatic disease treatment.[83] When this shadow cost of treating adverse effects is taken into account, and assuming equal efficacy (which remains to be proven), the cost-benefit analysis shifts decidedly in favour of topical over oral NSAIDs.[64] It remains to be seen if the new COX-2specific NSAIDs can indeed reduce the GI morbidity and mortality associated with oral NSAID treatment of rheumatic disorders. 6. Conclusions Topical NSAIDs have been used to treat a wide variety of conditions: musculoskeletal injuries,[3,4] postoperative pain,[84,85] herpetic neuralgia,[86] periodontitis,[87] aphthous ulcers[88] and actinic keratoses.[89] The anti-inflammatory efficacy of topically applied NSAIDs, even remote from the site of application, has been well established in animal models.[90] Evidence for anti-inflammatory activity from topical NSAIDs in humans is more equivocal, characterised by a considerable placebo effect (5 to 60% in chronic rheumatic diseases[91]) and significant interpatient variability in response.
Adis International Limited. All rights reserved.

Local enhanced topical delivery of topically applied NSAIDs does occur in humans, but the tissue depth at which the systemic circulation takes over distribution of the drug is highly variable among individuals. Individual variability in subcutaneous vasculature may account for the wide range of tissue depths reported after topical administration, as well as inconsistency in patient response to topical NSAIDs. Few generalisations can be made regarding the best drug or topical dosage form for enhancing cutaneous NSAID penetration and efficacy. Optimisation of vehicle formulation, pH and occlusion have all been documented to enhance penetration. Transdermal application gives rise to much lower plasma concentrations than oral administration. NSAID concentrations in the synovium subjacent to the site of topical application are at least comparable to those achieved after equivalent oral administration; subcutaneous concentrations after topical administration far exceed that which are normally achieved after oral administration. Adverse effects secondary to topical NSAID application occur in approximately 10 to 15% of patients, and are primarily cutaneous in nature (rash and pruritus at site of application). GI ADRs are rare with topically applied NSAIDs, compared with a 15% incidence reported for oral NSAIDs. Positive treatment outcomes reported in patients with chronic rheumatic disorders range from 30 to 95%, with considerable interpatient variability.[91] On average, 1 out of 3 patients using topical NSAIDs will achieve a successful outcome who would not have done so had they used a placebo.[91] The clinical evidence suggests, but does not prove, that NSAIDs applied topically are as effective as oral NSAIDs in the treatment of chronic rheumatic diseases. However, the contribution to efficacy by the rather dramatic placebo effect seen in these studies remains to be teased out by larger, well controlled, crossover studies comparing oral administration of clinically effective doses with topical formulations with proven superiority over vehicle controls. The safety profile of topical NSAIDs in terms of cutaneous ADRs is approximately
Drugs 2000 Sep; 60 (3)

572

Heyneman et al.

equivalent to that of oral formulations, but topical preparations are clearly superior in terms of GI adverse effects. Given the relatively benign adverse effect profile of topical NSAIDs and the increasing risk of serious NSAID-mediated GI events with age, it seems rational to give patients with rheumatic disorders a trial of topical NSAIDs prior to institution of an oral NSAID if oral paracetamol fails to control their pain. References
1. Grahame R. The NSAID patch [editorial]. Eur J Rheumatol Inflamm 1994; 14 (4): 1-2 2. Grahame R. Transdermal nonsteroidal anti-inflammatory agents. Br J Clin Pract 1995 Jan-Feb; 49 (1): 33-5 3. Vaile JH, Davis P. Topical NSAIDs for musculoskeletal conditions: a review of the literature. Drugs 1998 Nov; 56 (5): 783-99 4. Heyneman CA. Topical nonsteroidal anti-inflammatory drugs for acute soft tissue injuries. Ann Pharmacother 1995 Jul-Aug; 29 (7-8): 780-2 5. American College of Rheumatology Ad Hoc Committee on Clinical Guidelines. Guidelines for the management of rheumatoid arthritis. Arthritis Rheum 1996 May; 39 (5): 713-22 6. Schumacher HR. Classification of the rheumatic diseases. In: Schumacher Jr HR, editor. Primer on the rheumatic diseases. 10th ed. Atlanta (GA): Arthritis Foundation, 1993: 2-6 7. Paulus HE, Bulpitt KJ. Nonsteroidal anti-inflammatory agents and corticosteroids. In: Schumacher Jr HR, editor. Primer on the rheumatic diseases. 10th ed. Atlanta (GA): Arthritis Foundation, 1993: 298-303 8. Boh LE. Osteoarthritis. In: DiPiro JT, Talbert RL, Yee GC, et al., editors. Pharmacotherapy: a pathophysiological approach. 4th ed. Norwalk (CT): Appleton & Lange, 1999: 1441-59 9. Bagge E, Brooks P. Osteoarthritis in older patients: optimum treatment. Drugs Aging 1995 Sep; 7 (3): 176-83 10. Biundo JJ. Regional rheumatic pain syndromes. In: Schumacher Jr HR, editor. Primer on the rheumatic diseases. 10th ed. Atlanta (GA): Arthritis Foundation, 1993: 277-87 11. Eccles M, Freemantle N, Mason J. North of England evidence based guideline development project: summary guideline for non-steroidal anti-inflammatory drugs versus basic analgesia in treating the pain of degenerative arthritis. BMJ 1998 Aug 22; 317 (7157): 526-30 12. Schnitzer TJ. Osteoarthritis treatment update: minimizing pain while limiting patient risk. Postgrad Med 1993 Jan; 93 (1): 89-92, 95 13. Hochberg MC, Altman RD, Brandt KD, et al. Guidelines for the medical management of osteoarthritis: part II. Osteoarthritis of the knee. Arthritis Rheum 1995; 38: 1541-6 14. Vane JR. Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs. Nat New Biol 1971 Jun 23; 231 (25): 232-5 15. Brooks PM, Day RO. Nonsteroidal anti-inflammatory drugs: differences and similarities. N Engl J Med 1991 Jun 13; 324 (24): 1716-25 16. Gierse JK, Koboldt CM, Walker MC, et al. Kinetic basis for selective inhibition of cyclo-oxygenases. Biochem J 1999 May 1; 339 Pt 3: 607-14

17. Squire I, Lees K. Topical drug delivery. Practitioner 1992 Feb; 236: 203-6 18. Yano T, Nakagawa A, Tsuji M, et al. Skin permeability of various non-steroidal anti-inflammatory drugs in man. Life Sci 1986 Sep 22; 39 (12): 1043-50 19. Brown MB, Marriott C, Martin GP. The effect of hyaluronan on the in vitro deposition of diclofenac within the skin. Int J Tissue React 1995; 17 (4): 133-40 20. Rao PR, Diwan PV. Formulation and in vitro evaluation of polymeric films of diltiazem hydrochloride and indomethacin for transdermal administration. Drug Dev Ind Pharm 1998 Apr; 24 (4): 327-36 21. Tomida H, Shinohara M, Kuwada N, et al. In vitro release characteristics of diclofenac and hydrocortisone from Pluronic F127 gels. Acta Pharm Suec 1987; 24 (5): 263-72 22. Priborsky J, Takayama K, Obata Y. Influence of limonene and laurocapram on percutaneous absorption of nonsteroidal antiinflammatory drugs. Arzneimittelforschung 1992 Feb; 42 (2): 116-9 23. Singh SK, Durrani MJ, Reddy IK. Effect of permeation enhancers on the release of ketoprofen through transdermal drug delivery systems. Pharmazie 1996 Oct; 51 (10): 741-4 24. Friedman DI, Schwarz JS, Weisspapir M. Submicron emulsion vehicle for enhanced transdermal delivery of steroidal and nonsteroidal anti-inflammatory drugs. J Pharm Sci 1995 Mar; 84 (3): 324-9 25. Singh P, Anlike M, Smith GA, et al. Transdermal iontophoresis and solute penetration across excised human skin. J Pharm Sci 1995 Nov; 84 (11): 1342-6 26. Akhter SA, Barry BW. Absorption through human skin of ibuprofen and flurbiprofen: effect of dose variation, deposited drug films, occlusion and the penetration enhancer N-methyl2-pyrrolidone. J Pharm Pharmacol 1985 Jan; 37 (1): 27-37 27. McNeill SC, Potts RO, Francoeur ML. Local enhanced topical delivery (LETD) of drugs: does it truly exist? Pharm Res 1992 Nov; 9 (11): 1422-7 28. Mikulak SA, Vangsness CT, Nimni ME. Transdermal delivery and accumulation of indomethacin in subcutaneous tissues in rats. J Pharm Pharmacol 1998 Feb; 50 (2): 153-8 29. Singh P, Roberts MS. Skin permeability and local tissue concentrations of nonsteroidal anti-inflammatory drugs after topical application. J Pharmacol Exp Ther 1994 Jan; 268 (1): 144-51 30. Treffel P, Bagard B. Ibuprofen epidermal levels after topical application in vitro: effect of formulation, application time, dose variation and occlusion. Br J Dermatol 1993 Sep; 129 (3): 286-91 31. Cordero JA, Alarcon L, Escribano E, et al. A comparative study of the transdermal penetration of a series of nonsteroidal antiinflammatory drugs. J Pharm Sci 1997 Apr; 86 (4): 503-8 32. Roy SD, Manoukian E. Transdermal delivery of ketorolac tromethamine: permeation enhancement, device design, and pharmacokinetics in healthy humans. J Pharm Sci 1995 Oct; 84 (10): 1190-6 33. Seth BL. Comparative pharmacokinetics and bioavailability study of percutaneous absorption of diclofenac from two topical formulations containing drug as a solution gel or as an emulsion gel. Arzneimittelforschung 1992 Feb; 42 (2): 120-2 34. Seth PL. Percutaneous absorption of ibuprofen from different formulations: comparative study with gel, hydrophilic ointment and emulsion cream. Arzneimittelforschung 1993 Aug; 43 (8): 919-21

Adis International Limited. All rights reserved.

Drugs 2000 Sep; 60 (3)

Oral versus Topical NSAIDs in Rheumatic Diseases

573

35. Roy SD, Manoukian E, Combs D. Absorption of transdermally delivered ketorolac acid in humans. J Pharm Sci 1995 Jan; 84 (1): 49-52 36. Gevi M, Merlo M. Ketoprofen lysine by topical route in sports traumatology. Curr Ther Res 1983 Nov; 34 (5): 844-50 37. Kanazawa M, Ito H. The pharmacokinetics of 0.5% piroxicam gel in humans. Eur J Rheumatol Inflamm 1997; 8 (1): 117 38. Muller M, Rastelli C, Ferri P, et al. Transdermal penetration of diclofenac after multiple epicutaneous administration. J Rheumatol 1998 Sep; 25 (9): 1833-6 39. Assandri A, Canali S, Giachetti C. Local tolerability and pharmacokinetic profile of a new transdermal delivery system, diclofenac hydroxyethylpyrrolidine plaster. Drugs Exp Clin Res 1993; 19 (3): 89-95 40. Shah AK, Wei G, Lanman RC, et al. Percutaneous absorption of ketoprofen from different anatomical sites in man. Pharm Res 1996 Jan; 13 (1): 168-72 41. Sioufi A, Pommier F, Boschet F, et al. Percutaneous absorption of diclofenac in healthy volunteers after single and repeated topical application of diclofenac Emulgel. Biopharm Drug Dispos 1994 Aug; 15 (6): 441-9 42. Taburet AM, Singlas E, Glass RC, et al. Pharmacokinetic comparison of oral and local action transcutaneous flurbiprofen in healthy volunteers. J Clin Pharm Ther 1995 Apr; 20 (2): 101-7 43. Flouvat B, Roux A, Delhotal-Landes B. Pharmacokinetics of ketoprofen in man after repeated percutaneous administration. Arzneimittelforschung 1989 Jul; 39 (7): 812-5 44. Tegeder I, Muth-Selbach U, Lotsch J, et al. Application of microdialysis for the determination of muscle and subcutaneous tissue concentrations after oral and topical ibuprofen administration. Clin Pharmacol Ther 1999 Apr; 65 (4): 357-68 45. Dominkus M, Nicolakis M, Kotz R, et al. Comparison of tissue and plasma levels of ibuprofen after oral and topical administration. Arzneimittelforschung 1996 Dec; 46 (12): 1138-43 46. McNeil Consumer Products Co. Motrin (ibuprofen) [package insert - revised]. Fort Washington (PA): McNeil Consumer Products Co., Dec 1994 47. Rolf C, Engstrom B, Beauchard C, et al. Intra-articular absorption and distribution of ketoprofen after topical plaster application and oral intake in 100 patients undergoing knee arthroscopy. Rheumatology (Oxford) 1999 Jun; 38 (6): 564-7 48. Cushman DW, Cheung HS. Effect of substrate concentration on inhibition of prostaglandin synthetase of bull seminal vesicles by anti-inflammatory drugs and fenamic acid analogs. Biochim Biophys Acta 1976 Mar 26; 424 (3): 449-59 49. Gallacchi G, Marcolongo R. Pharmacokinetics of diclofenac hydroxyethylpyrrolidine (DHEP) plasters in patients with monolateral knee joint effusion. Drugs Exp Clin Res 1993; 19 (3): 95-7 50. Day RO, McLachlan AJ, Graham GG, et al. Pharmacokinetics of nonsteroidal anti-inflammatory drugs in synovial fluid. Clin Pharmacokinet 1999 Mar; 36 (3): 191-210 51. Mattara L, Trotta F, Biasi D, et al. Evaluation of the efficacy and tolerability of a new locally acting preparation of flurbiprofen in scapulohumeral periarthritis. Eur J Rheumatol Inflamm 1994; 14 (4): 15-20 52. Poul J, West J, Buchanan N, et al. Local action transcutaneous flurbiprofen in the treatment of soft tissue rheumatism. Br J Rheumatol 1993 Nov; 32 (11): 1000-3 53. Saggini R, Zoppi M, Vecchiet F, et al. Comparison of electromotive drug administration with ketorolac or with placebo in patients with pain from rheumatic disease: a double-masked study. Clin Ther 1996 Nov-Dec; 18 (6): 1169-74

54. Shackel NA, Day RO, Kellett B, et al. Copper-salicylate gel for pain relief in osteoarthritis: a randomised controlled trial. Med J Aust 1997 Aug 4; 167 (3): 134-6 55. Dreiser RL, Tisne-Camus M. DHEP plasters as a topical treatment of knee osteoarthritis: a double-blind placebo-controlled study. Drugs Exp Clin Res 1993; 19 (3): 117-23 56. Galeazzi M, Marcolongo R. A placebo-controlled study of the efficacy and tolerability of a nonsteroidal anti-inflammatory drug, DHEP plaster, in inflammatory peri- and extra-articular rheumatological diseases. Drugs Exp Clin Res 1993; 19 (3): 107-15 57. Roth SH. A controlled clinical investigation of 3% diclofenac/ 2.5% sodium hyaluronate topical gel in the treatment of uncontrolled pain in chronic oral NSAID users with osteoarthritis. Int J Tissue React 1995; 17 (4): 129-32 58. Ginsberg F, Famaey JP. Double-blind, randomized crossover study of the percutaneous efficacy and tolerability of a topical indomethacin spray versus placebo in the treatment of tendinitis. J Int Med Res 1991 Mar-Apr; 19 (2): 131-6 59. Grace D, Rogers J, Skeith K, et al. Topical diclofenac versus placebo: a double blind randomized clinical trial in patients with osteoarthritis of the knee. J Rheumatol 1999; 26: 2659-63 60. Waikakukl S, Danputipong P, Soparat K. Topical analgesics, indomethacin plaster and diclofenac emulgel for low back pain: a parallel study. J Med Assoc Thai 1996 Aug; 79 (8): 486-90 61. Ritchie LD. A clinical evaluation of flurbiprofen LAT and piroxicam gel: a multicentre study in general practice. Clin Rheumatol 1996 May; 15 (3): 243-7 62. Rosenthal M, Bahous I. A controlled clinical study on the new topical dosage form of DHEP plasters in patients suffering from localized inflammatory diseases. Drugs Exp Clin Res 1993; 19 (3): 99-105 63. Giacovazzo M. Clinical evaluation of a new NSAID applied topically (BPAA gel) vs. diclofenac emulgel in elderly osteoarthritic patients. Drugs Exp Clin Res 1992; 18 (5): 201-3 64. Hosie G, Bird H. The topical NSAID felbinac versus oral NSAIDS: a critical review. Eur J Rheumatol Inflamm 1994; 14 (4): 21-8 65. Tsuyama N, Kurokawa T, Hihei T, et al. Clinical evaluation of L-141 topical agent on osteoarthrosis deformans of the knees. Clin Med 1985; 1: 697-729 66. Martens M. Efficacy and tolerability of a topical NSAID patch (local action transcutaneous flurbiprofen) and oral diclofenac in the treatment of soft-tissue rheumatism. Clin Rheumatol 1997 Jan; 16 (1): 25-31 67. Sandelin J, Harilainen A, Crone H, et al. Local NSAID gel (eltenac) in the treatment of osteoarthritis of the knee: a double blind study comparing eltenac with oral diclofenac and placebo gel. Scand J Rheumatol 1997; 26 (4): 287-92 68. Dickson DJ. A double-blind evaluation of topical piroxicam gel with oral ibuprofen in osteoarthritis of the knee. Curr Ther Res 1991 Feb; 49 (2): 199-207 69. Browning RC, Johson K. Reducing the dose of oral NSAIDs by use of Feldene Gel: an open study in elderly patients with osteoarthritis. Adv Ther 1994 Jul-Aug; 11 (4): 198-207 70. Evans JM, MacDonald TM. Tolerability of topical NSAIDs in the elderly: do they really convey a safety advantage? Drugs Aging 1996 Aug; 9 (2): 101-8 71. Wynne HA, Rawlins MD. Are systemic levels of nonsteroidal anti inflammatory drugs relevant to acute upper gastrointestinal haemorrhage? Eur J Clin Pharmacol 1993; 44 (4): 309-13

Adis International Limited. All rights reserved.

Drugs 2000 Sep; 60 (3)

574

Heyneman et al.

72. Evans JMM, McGregor E, McMahon AD, et al. Non-steroidal anti-inflammatory drugs and hospitalization for acute renal failure. Q J Med 1995 Aug; 88 (8): 551-7 73. Carson JL, Strom BL, Soper KA, et al. The association of nonsteroidal anti-inflammatory drugs with upper gastrointestinal tract bleeding. Arch Intern Med 1987 Jan; 147: 85-8 74. Figueras A, Capella D, Castel JM, et al. Spontaneous reporting of adverse drug reactions to non-steroidal anti-inflammatory drugs: a report from the Spanish System of Pharmacovigilance, including an early analysis of topical and enteric-coated formulations. Eur J Clin Pharmacol 1994; 47 (4): 297-303 75. Evans JMM, McMahon AD, McGilchrist MM, et al. Topical non-steroidal anti-inflammatory drugs and admission to hospital for upper gastrointestinal bleeding and perforation: a record linkage case-control study. BMJ 1995 Jul 1; 311 (6996): 22-6 76. Ophaswongse S, Maibach H. Topical nonsteroidal antiinflammatory drugs: allergic and photoallergic contact dermatitis and phototoxicity. Contact Dermatitis 1993 Aug; 29 (2): 57-64 77. Singh G, Ramey DR, Morfeld D, et al. Comparative toxicity of non-steroidal anti-inflammatory agents. Pharmacol Ther 1994 Apr-May; 62 (1-2): 175-91 78. Singh G, Ramey DR, Morfeld D, et al. Gastrointestinal tract complications of nonsteroidal anti-inflammatory drug treatment in rheumatoid arthritis: a prospective observational cohort study. Arch Intern Med 1996 Jul 22; 156 (14): 1530-6 79. Vestal RE. Geriatric clinical pharmacology: an overview. In: Vestal RE, editor. Drug therapy in the elderly. Sydney: ADIS Health Science Press, 1984: 12-28 80. Joint Working Group of the British Society for Rheumatology and the Research Unit of the Royal College of Physicians: guidelines for the diagnosis, investigation and management of osteoarthritis of the hip and knee. J R Coll Physicians Lond 1993; 27: 391-6 81. Gabriel SE, Jaakkimainen L, Bombardier C. Risk for serious gastrointestinal complications related to use of nonsteroidal anti-inflammatory drugs: a meta-analysis. Ann Intern Med 1991 Nov 15; 115 (10): 787-96 82. Topical NSAIDs: a gimmick or a godsend [editorial]? Lancet 1989 Sep 30; 2 (8666): 779-80

83. Bloom BS. Direct medical costs of disease and gastrointestinal side effects during treatment for arthritis. Am J Med 1988 Feb 22; 84 (2A): 20-4 84. Liu K, Chia Y-Y, Lo Y, et al. Effect of preoperative transdermal ketoprofen on post-hysterectomy pain. Chung Hua I Hsueh Tsa Chih (Taipei) 1997 Dec; 60 (6): 290-5 85. Reinhart DJ, Goldberb MI, Roth JV, et al. Transdermal fentanyl system plus im ketorolac for the treatment of postoperative pain. Can J Anaesth 1997 Apr; 44 (4): 377-84 86. De Benedittis G, Lorenzetti A. Topical aspirin/diethyl ether mixture versus indomethacin and diclofenac/diethyl ether mixtures for acute herpetic neuralgia and postherpetic neuralgia: a double-blind crossover placebo-controlled study. Pain 1996 Apr; 65 (1): 45-51 87. Lawrence HP, Paquette DW, Smith PC, et al. Pharmacokinetic and safety evaluations of ketoprofen gels in subjects with adult periodontitis. J Dent Res 1998 Nov; 77 (11): 1904-12 88. Saxen MA, Ambrosius WR, Rehemtula al-KF, et al. Sustained relief of oral aphthous ulcer pain from topical diclofenac in hyaluronan: a randomized, double-blind clinical trial. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1997 Oct; 84 (4): 356-61 89. Rivers JK. An open study to assess the efficacy and safety of topical 3% diclofenac in a 2.5% hyaluronic acid gel for the treatment of actinic keratoses. Arch Dermatol 1997 Oct; 133 (10): 1239-42 90. Hiramatsu Y, Akita S, Salamin PA, et al. Assessment of topical non-steroidal anti-inflammatory drugs in animal models. Arzneimittelforschung 1990 Oct; 40 (10): 1117-24 91. Moore RA, Tramer MR, Carroll D. Quantitative systematic review of topically applied non-steroidal anti-inflammatory drugs. BMJ 1998 Jan 31; 316 (7128): 333-8

Correspondence and offprints: Dr Catherine A. Heyneman, Assistant Professor of Pharmacy Practice, Idaho State University College of Pharmacy, Campus Box 8356, Pocatello, ID 83209, USA. E-mail: cathy@otc.isu.edu

Adis International Limited. All rights reserved.

Drugs 2000 Sep; 60 (3)