Clinical and Pharmacogenetic Influences on Response to ... - Core

See related commentary on pg 1968

ORIGINAL ARTICLE

Clinical and Pharmacogenetic Influences on Response to Hydroxychloroquine in Discoid Lupus Erythematosus: A Retrospective Cohort Study Shyamal Wahie1, Ann K. Daly1, Heather J. Cordell2, Mark J. Goodfield3, Stephen K. Jones4, Christopher R. Lovell5, Andrew J. Carmichael6, Mary M. Carr7, Angela Drummond8, Sivakumar Natarajan9, Catherine H. Smith10, Nick J. Reynolds1,11 and Simon J. Meggitt1,11 The recommended systemic therapy of choice for discoid lupus erythematosus (DLE) is the 4-aminoquinolone antimalarial hydroxychloroquine. There is limited published information on the likelihood of clinical response and, in particular, what factors influence outcome. We conducted a multicenter observational and pharmacogenetic study of 200 patients with DLE treated with hydroxychloroquine. The primary outcome was clinical response to hydroxychloroquine. We investigated the effects of disease attributes and metabolizing cytochrome P450 (CYP) polymorphisms on clinical outcome. Although the majority of patients responded to hydroxychloroquine, a significant proportion (39%) either failed to respond or was intolerant of the drug. Cigarette smoking and CYP genotype did not have any significant influence on response to hydroxychloroquine. Moreover, multivariate analysis indicated that disseminated disease (odds ratio (OR): 0.21; 95% confidence interval (CI): 0.08–0.52; Po0.001) and concomitant systemic lupus erythematosus (SLE; OR: 0.06; 95% CI: 0.01–0.49; P ¼ 0.009) were significantly associated with lack of response to hydroxychloroquine. These findings suggest that baseline lupus severity and SLE are predictors of response to hydroxychloroquine. A prospective study is now required to further investigate the relationship between disease activity and response to hydroxychloroquine. This will have the potential to further inform the clinical management of this disfiguring photosensitive disease. Journal of Investigative Dermatology (2011) 131, 1981–1986; doi:10.1038/jid.2011.167; published online 7 July 2011

INTRODUCTION Discoid lupus erythematosus (DLE) is the most common form of cutaneous lupus, (Tebbe and Orfanos, 1997; Kyriakis et al., 2007) and can cause irreversible scarring and disfigurement. Hydroxychloroquine is the oral therapy of choice (Jessop et al., 2009; Walling and Sontheimer, 2009). The factors influencing clinical response to hydroxychloroquine are 1

Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne, UK; 2Institute of Genetic Medicine, Newcastle University, Newcastle upon Tyne, UK; 3Department of Dermatology, Leeds General Infirmary, Leeds, UK; 4 Department of Dermatology, Clatterbridge Hospital, Wirral, UK; 5 Department of Dermatology, Royal United Hospital Bath, Bath, UK; 6 Department of Dermatology, James Cook University Hospital, Middlesbrough, UK; 7Department of Dermatology, University Hospital of North Durham, Durham, UK; 8Department of Dermatology, Southern General Hospital, Glasgow, UK; 9Department of Dermatology, Sunderland Royal Hospital, Sunderland, UK; 10St John’s Institute of Dermatology, St Thomas’ Hospital, London, UK and 11Department of Dermatology, Royal Victoria Infirmary, Newcastle upon Tyne, UK Correspondence: Shyamal Wahie, School of Clinical and Laboratory Sciences, Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne NE2 4HH, UK. E-mail: [email protected] Abbreviations: CI, confidence interval; DLE, discoid lupus erythematosus; OR, odds ratio; SLE, systemic lupus erythematosus Received 24 September 2010; revised 11 March 2011; accepted 1 April 2011; published online 7 July 2011

& 2011 The Society for Investigative Dermatology

unclear. Small single-site retrospective studies (Rahman et al., 1998; Jewell and McCauliffe, 2000) have suggested that cigarette smoking is associated with poor response to hydroxychloroquine, but this has not been confirmed prospectively (Lardet et al., 2004) and potential confounding variables have not been assessed. Nevertheless, this possible association has come to be a widely accepted ‘‘fact’’ in dermatology literature and clinics alike (Kuhn et al., 2010). Patients with systemic lupus erythematosus (SLE) and rheumatoid arthritis, treated with hydroxychloroquine, demonstrate a positive correlation between whole-blood hydroxychloroquine levels and clinical response (Tett et al., 1993, 2000; Munster et al., 2002; Costedoat-Chalumeau et al., 2006). It is plausible that variation in therapeutic response to hydroxychloroquine might, in part, be due to inherent differences in hepatic metabolism of hydroxychloroquine between individuals, which therein influence whole-blood hydroxychloroquine levels. Chloroquine is metabolized by the cytochrome P450 (CYP) enzymes CYP2C8, CYP3A4, and CYP2D6 (Kim et al., 2003; Projean et al., 2003), with CYP2C8 contributing the most. As hydroxychloroquine and chloroquine differ in structure by only one hydroxyl group, it is very likely that hydroxychloroquine is also metabolized by the same enzymes. www.jidonline.org 1981

S Wahie et al. Response to Hydroxychloroquine in DLE

Few functionally significant polymorphisms of CYP3A4 have been identified and are extremely rare in the United Kingdom. Although it is not known whether polymorphisms in CYP2C8 and 2D6 account for part of the differential response of DLE to hydroxychloroquine, the metabolism of the antimalarial amodiaquine (structurally related to chloroquine) appears to be impaired in individuals with CYP2C8 polymorphisms (Gil and Gil Bergluand, 2007; Parikh et al., 2007). If CYP2C8 and/or CYP2D6 polymorphisms are associated with impaired metabolism of hydroxychloroquine, individuals possessing these variant alleles might respond better to hydroxychloroquine than those that are wild type (via an indirect effect on hydroxychloroquine levels). However, the importance of such effects on response in comparison with host or disease attributes (such as smoking status or disease extent) is unknown. We conducted a multicenter observational and pharmacogenetic cohort study in 200 patients with DLE treated with hydroxychloroquine. The outcomes of interest were: (i) the clinical response to hydroxychloroquine and (ii) the effects of disease attributes, smoking, and metabolizing CYP polymorphisms on clinical response. RESULTS The median age at diagnosis was 40 years (range 16–81), with a female to male ratio of 4:1. Median follow-up was for 8 years (range 0.5–37). In all, 92% of patients were seen in standard dermatology clinics (spanning both district general and teaching hospitals) and 8% were ‘‘tertiary referrals’’ (i.e., referrals sent from one dermatologist to another for further management). A total of 11 patients (5.5%) had SLE; 8 had inflammatory arthritis; 5 had nephritis, including 1 requiring renal transplant; 2 had mouth ulcers; and 1 had pulmonary fibrosis. CYP2C8 and 2D6 polymorphisms

Raw allele numbers (n) and allele frequencies (f) were consistent with those expected amongst Caucasians. For CYP2D6*1: n ¼ 300, (f) 0.75; *3: n ¼ 3, (f) 0.01; *4: n ¼ 83, (f) 0.21; and *5: n ¼ 14, (f) 0.04. For CYP2C8*1: n ¼ 324, (f) 0.81; *3: n ¼ 50, (f) 0.13; and *4: n ¼ 26, (f) 0.07. Response within 6 months

Response to hydroxychloroquine occurred in 120 patients (60%). A total of 70 patients (35%) were non-responders, and in 8 patients (4%), therapy was withdrawn early because of adverse effects (three with gastrointestinal events, three with rashes, one with malaise, and one with palpitations) before response could be assessed. In two patients (1%), response to hydroxychloroquine alone was equivocal because it was commenced in conjunction with mepacrine from the outset. Clinical characteristics and outcome to therapy

Table 1 indicates the clinical characteristics of the patients and the results of univariate logistic regression analysis. All 200 patients satisfied our inclusion criteria for a previous histological and clinical diagnosis of DLE. At the time of recruitment, there were two patients who had some clinical 1982 Journal of Investigative Dermatology (2011), Volume 131

signs of subacute cutaneous lupus in addition to their DLE. These overlap features had only emerged years after receiving hydroxychloroquine for DLE and were still not the predominant feature of their disease. As both patients received hydroxychloroquine at a time when they only had DLE, a response to therapy could still be assigned for that period in time. These two patients, who were both responders to hydroxychloroquine, were not excluded from the analysis. The patients who had equivocal response (n ¼ 2) or withdrew from therapy because of side effects before response being assignable (n ¼ 8) were excluded from analysis. Patients with localized disease were significantly more likely to respond compared with those who had disseminated disease (odds ratio (OR): 7.29; 95% confidence interval (CI): 3.16–16.79; Po0.0001). There was a significant trend for increasing number of body sites affected with DLE to be associated with reduced probability of response (test for trend; P ¼ 1.78 106). In addition, those patients without associated features of SLE were also significantly more likely to respond in contrast to those with SLE (OR: 19.83; 95% CI: 2.48–159; P ¼ 0.005). During the first 6 months of hydroxychloroquine use, 75 patients (38%) did not smoke, of whom 43 (57%) responded and 29 (39%) did not. A total 125 patients (63%) were regular smokers, of whom 77 (62%) responded and 41 (33%) did not. During the first 6 months of hydroxychloroquine use, there was no significant difference in the response rate between non-smokers compared with smokers (OR: 0.78; 95% CI: 0.43–1.45; P ¼ 0.27). Table 2 shows the cohort’s CYP genotypes alongside their response. There was a nonsignificant trend for CYP2C8 and 2D6 variants to be associated with response to hydroxychloroquine. However, the ORs detected are in the range of 1.20–1.55 (with wide CIs), in contrast to some of the significant effect sizes demonstrated in Table 1. Table 3 shows the results of the multivariate logistic regression analysis. Localized disease remained significantly associated with response to hydroxychloroquine (OR: 4.80; 95% CI: 1.93–11.97; P ¼ 0.001). In contrast, DLE with concomitant SLE was significantly associated with low odds of response to hydroxychloroquine (OR: 0.06; 95% CI: 0.01–0.49; P ¼ 0.009). This refers to the lack of response of DLE itself to therapy and not to the response of the other manifestations of SLE. Tertiary referrals were also associated with poor response to hydroxychloroquine (OR: 0.16; 95% CI: 0.04–0.66; P ¼ 0.011). All tertiary referrals were seen at one center. After exclusion of this particular tertiary referral center, there was no significant difference amongst the remaining seven centers within the response rates (P ¼ 0.36). After excluding all tertiary referrals from the analysis, 35% of the remaining 183 patients still either failed to respond to or withdrew from hydroxychloroquine and the factors associated with response in the univariate and multivariate analyses remained the same. DISCUSSION This is the largest study of patients with DLE investigated for response to hydroxychloroquine. Although the majority


Table 1. Univariate logistic regression analysis testing association between clinical factors and response Nonresponder

Responder

M

17

23

F

53

97

1

20

65

2

20

40

3

4

6

4

14

5

5

11

4

Variable

OR (95% CI) for response

P-value

0.73 (0.36–1.50)

0.26

0.538 per site (0.42–0.69)

1.78 106

Localized versus disseminated

7.29 (3.16–16.79)

o0.0001

Photodistributed only versus non-photodistributed

3.13 (1.59–6.19)

0.001

No scarring alopecia versus scarring alopecia

2.13 (1.17–3.88)

0.014

Non-smoker versus smoker

0.78 (0.43–1.45)

0.27

bd versus od

0.58 (0.28–1.19)

0.09

ve versus +ve

1.87 (0.97–3.63)

0.062

SLE versus no SLE

0.05 (0.01–0.40)

0.005

Tertiary versus non-tertiary

0.10 (0.03–0.37)

o0.0001

Test M versus F

Sites affected

1

0

Localized

6

44

111

Disseminated

26

9

Sites affected include all photodistributed areas

43

100

Sites affected include non-photodistributed areas

27

20

No scarring alopecia

32

77

Scarring alopecia

38

43

Non-smoker

29

43

Smoker

41

77

200 mg od dose

13

32

200 mg bd dose

57

81

200 mg tds dose

0

7

ANF ve or borderline/ENA/dsDNA ve

34

76

ANF +ve and/or ENA +ve/borderline and/or dsDNA +ve/borderline

26

31

With SLE

10

1

Without SLE

60

119

Tertiary

14

3

Non-tertiary

56

117

Test for trend

Test for trend

0.26

Abbreviations: ANF, anti-nuclear factor; bd, twice daily; CI, confidence interval; dsDNA, double-stranded DNA; ENA, extractable nuclear antigen; F, female; M, male; od, once daily; OR, odds ratio; SLE, systemic lupus erythematosus; tds, three times daily.

responded to hydroxychloroquine, a significant proportion (39%) either failed to respond to the drug or developed toxicity, requiring withdrawal of therapy. This emphasizes the need to investigate the reasons for poor response to therapy. Clinical characteristics of the cohort were similar to those reported by others (Vera-Recabarren et al., 2010). Multivariate analysis showed that localized disease (involving three or less body sites) and absence of concomitant SLE were significantly associated with response to hydroxychloroquine. We found that individuals with disseminated DLE (including disease in non-photodistributed sites and/or scarring alopecia), and/or features indicating current or emerging SLE, were unlikely to respond to hydroxychloroquine. In these patients, close monitoring might be prudent

and other therapies might need to be considered early during the course of their disease in an attempt to avoid disfiguring scarring. A recent single-center cross-sectional study of 114 patients with cutaneous lupus indicated that treatment-refractory patients were more likely to have generalized disease (Moghadam-Kia et al., 2009). Whether generalized disease was a consequence of poor response to therapy or had an influence on outcome to therapy is difficult to discern, as the majority of patients assessed were already receiving established medical care when included for analysis. In contrast, our study has reported the clinical characteristics of patients before hydroxychloroquine being commenced and shown that disease extent does influence eventual response. www.jidonline.org 1983


Table 2. Univariate logistic regression analysis testing associations between CYP genotype and response CYP gene

Genotype

2C8

*1/*1 (WT) *1/*3 *3/*4

1

1

*3/*3

2

7

2D6

Non-responder

Responder

Test


P-value

51

76

All mutants versus WT

1.55 (0.82–2.96)

0.18

10

18

All mutants versus WT

1.20 (0.67–2.18)

0.54

*1/*4

6

18

*1/*1 (WT)

40

63

*1/*4

21

42

*1/*3

1

1

*4/*4

4

3

*3/*4

0

1

*1/*5

4

8

*4/*5

0

2

Abbreviations: CI, confidence interval; CYP, cytochrome P450; OR, odds ratio; WT, wild type.

Table 3. Multivariate logistic regression analysis showing factors associated with response Variable


P-value

Localized disease

4.80 (1.93–11.97)

0.001

SLE

0.06 (0.01–0.49)

0.009

Tertiary referral

0.16 (0.04–0.66)

0.011

Abbreviations: CI, confidence interval; OR, odds ratio; SLE, systemic lupus erythematosus.

Our study did not demonstrate a significant association between CYP genotype and response, although there was a trend for CYP2C8 variants to be associated with greater odds of response. These variants have previously been demonstrated to be associated with slow metabolism of several drugs (e.g., Parikh et al., 2007; Niemi et al., 2003), although their precise effect on hydroxychloroquine has not been examined. If the true effect of CYP polymorphisms on chance of response approximates to an OR of 1.55 as we have found, further power calculations indicate that an extra 500 patients would be required to establish this association (at 1% significance level and maintaining 80% power). Previous studies have indicated that B80% of patients with cutaneous lupus are smokers (Miot et al., 2005; Boeckler et al., 2009) and our data is consistent with this. It has been suggested that cigarette smoking might influence the response of DLE to antimalarials, but studies have been contradictory (Rahman et al., 1998; Jewell and McCauliffe, 2000; Lardet et al., 2004). We have demonstrated no significant difference in response rates between smokers and non-smokers, casting doubt on the notion that smoking per se inhibits response to antimalarials in DLE. Unlike other studies (Rahman et al., 1998; Jewell and McCauliffe, 2000), we controlled for 1984 Journal of Investigative Dermatology (2011), Volume 131

confounding variables and our cohort is at least four times larger than others. This finding marks an important reversal of dermatology dogma. Although there are limitations in undertaking a retrospective assessment of response, we attempted to minimize the potential for bias with the following measures: a single investigator recruited and interviewed all patients, reviewed their case notes according to a predefined protocol, and importantly used a comprehensive, prespecified adjective list to designate response. We have since developed a scoring instrument for measuring DLE activity, which can be used prospectively to provide more robust studies of disease response and effect of CYP genotype (Wahie et al., 2010). In summary, we have identified that at least one in three patients with DLE fail to respond to the hydroxychloroquine within 6 months. Both disseminated disease and SLE at baseline are associated with poor outcome. In contrast, CYP genotype and smoking have no significant effect on response. A prospective study is now required to further test whether baseline disease severity and associated SLE predict response to hydroxychloroquine therapy. If this is confirmed, future studies will need to concentrate on the potential factors influencing disease activity. These might include effects of disease susceptibility genes and proinflammatory cytokines. MATERIALS AND METHODS In order to be included into the study, patients needed to have: (i) a clinical diagnosis of DLE with features of scarring, erythematous, and/or scaly plaques; (ii) a lesional skin biopsy demonstrating a periadnexal and/or perivascular chronic inflammatory infiltrate and epidermal ‘‘lichenoid’’ interface change; and (iii) previous treatment with hydroxychloroquine. Power calculations indicated that with 200 patients, the study would have 80% power to detect an OR of between 2 and 2.5 between non-responders and responders with respect to the CYP


genotype frequencies (at 1% significance level). Patients were recruited from eight centers and interviewed in person by a single investigator. The study received regulatory approvals from the Multicentre Research Ethics Committee (MREC) and the Medicines and Healthcare Products Regulatory Agency (MHRA) and was conducted according to the Declaration of Helsinki Principles. Written informed consent was obtained from all patients. Case notes were reviewed by the same investigator for: age at diagnosis; body sites affected by DLE at the clinic assessment immediately before commencing hydroxychloroquine, the ‘‘photodistributed’’ sites were the scalp, face, neck, arms, and hands, and the ‘‘non-photodistributed’’ sites were the torso, legs, and feet; disease extent (‘‘localized disease’’ was defined as DLE in three or less body sites and ‘‘disseminated disease’’ as DLE in four or more body sites); presence or absence of scarring alopecia; features of concomitant SLE (Hochberg, 1997); drug, alcohol, and smoking history (Ghaussy et al., 2003); disease duration; starting dose of hydroxychloroquine; height and weight; side effects; length of hospital follow-up; presence of antinuclear antibody (positive if 41:40 titer); precipitating antibodies to extractable nuclear antigens (Ro/SS-A and La/SS-B); and anti double-stranded DNA antibody (normal range 0–25 U ml1). An assessment of response to hydroxychloroquine by 6 months of use was informed by a retrospective case note reviewed by a single investigator. A predefined list was formulated containing 20 commonly used adjectives and phrases, which were likely to be used in medical case notes, to formalize retrospective designation of response. Examples of adjectives or phrases used to denote response by 6 months included: ‘‘better’’, ‘‘good’’, ‘‘significant improvement’’, ‘‘improved’’, ‘‘melted away’’, ‘‘controlled’’, ‘‘all clear’’, ‘‘virtually clear’’, ‘‘some improvement’’, and ‘‘lesions going/gone.’’ Examples of adjectives or phrases used to indicate non-response by 6 months included: ‘‘no change’’, ‘‘same’’, ‘‘not improving’’, ‘‘worse’’, ‘‘no better’’, ‘‘new lesions arising’’, ‘‘therapy needs changing’’, and ‘‘no significant change’’. All physician notes indicating patient progress to therapy during the first 6 months were documented to support the investigator’s final classification of outcome. Utilizing this approach, patients were classified as responders by 6 months if there had been an improvement in their disease and no other oral therapies were required to achieve that improvement. Patients were deemed to be non-responders by 6 months if their disease did not change or had worsened and/ or other oral therapies had been needed as a consequence. This definition of response is similar to that used in other retrospective and prospective studies assessing response to antimalarials in the context of cigarette smoking (Jewell and McCauliffe, 2000; Lardet et al., 2004). All patients donated whole blood (10 ml), and genomic DNA was isolated (Daly et al., 1996). Samples were genotyped using PCRRFLP for CYP2D6*3, *4, and *5 (Daly et al., 1996) and for CYP2C8*3 and *4 (Bahadur et al., 2002; Niemi et al., 2003).

Statistical analysis Univariate logistic regression analysis was performed (using STATA, Timberlake, London) to determine whether the CYP genotypes and clinical variables were associated with response. For each of the CYP genes of interest (2C8 and 2D6), response of wild-type (*1/*1) individuals was compared with heterozygotes and homozygous

mutants combined. Multivariate logistic regression analysis was undertaken putting all significant variables from the initial univariate analysis into a model and retaining them if they remained significantly associated with response (after adjustment for other factors) using a forward stepwise approach. CONFLICT OF INTEREST This study was supported and funded by Lupus UK. CHS undertakes research, which is partly funded by Abbott, Schering Plough, Pfizer, and Janssen Cilag. However, these grants are not used for any research associated with DLE or hydroxychloroquine. NJR’s research laboratory receives funding from AstraZeneca and Stefiel, a GlaxoSmithKline company. However, these grants are not used for any research associated with DLE or hydroxychloroquine. The remaining authors state no conflict of interest.

ACKNOWLEDGMENTS We acknowledge the following centers for identifying patients: Royal Victoria Infirmary, Leeds General Infirmary, Clatterbridge Hospital, Royal United Hospital Bath, James Cook University Hospital, University Hospital North Durham, Southern General Hospital, Sunderland Royal Hospital, and St Thomas’ Hospital. We thank the National Institute for Health Research (NIHR) Clinical Research Network for support. NJR’s laboratory/research is also supported by the Newcastle NIHR Biomedical Research Centre.

AUTHOR CONTRIBUTIONS The Discoid Lupus Research Group consists of: (i) SW, AKD, HJC, NJR, and SJM—these individuals provided substantial contributions to conception and design, acquisition of data, analysis and interpretation of data, and drafting the article, and they accept direct responsibility for the manuscript—(ii) MJG, SKJ, CRL, AJC, MMC, AD, SN, and CHS—all these individuals were participating investigators and contributed to identifying patients for the study.

REFERENCES Bahadur N, Leathart JB, Mutch E et al. (2002) CYP2C8 polymorphisms in Caucasians and their relationship with paclitaxel 6 alpha-hydroxylase activity in human liver microsomes. Biochem Pharmacol 64: 1579–89 Boeckler P, Cosnes A, France`s C et al. (2009) Association of cigarette smoking but not alcohol consumption with cutaneous lupus erythematosus. Arch Dermatol 145:1012–6 Costedoat-Chalumeau N, Amoura Z, Hulot JS et al. (2006) Low blood concentration of hydroxychloroquine is a marker for and predictor of disease exacerbations in patients with systemic lupus erythematosus. Arthritis Rheum 54:3284–90 Daly AK, Steen VM, Fairbrother KS et al. (1996) CYP2D6 multiallelism. Methods Enzymol 272:199–210 Gil JP, Gil Bergluand E (2007) CYP2C8 and antimalarial drug efficacy. Pharmacogenomics 8:187–98 Ghaussy WO, Sibbitt WL, Bankhurst AD et al. (2003) Cigarette smoking and disease activity in systemic lupus erythematosus. J Rheumatol 30:1215–21 Hochberg MC (1997) Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 40:1725 Jessop S, Whitelaw DA, Delamere FM (2009) Drugs for discoid lupus erythematosus. Cochrane Database Syst Rev 7:CD002954 Jewell JD, McCauliffe DP (2000) Patients with cutaneous lupus erythematosus who smoke are less responsive to antimalarial treatment. J Am Acad Dermatol 42:983–7 Kim KA, Park JY, Lee SJ et al. (2003) Cytochrome P450 2C8 and CYP 3A4/5 are involved in chloroquine metabolism in human liver microsomes. Arch Pharm Res 26:631–7 Kuhn A, Ruland V, Bonsmann G (2010) Cutaneous lupus erythematosus: update of therapeutic options part I. J Am Acad Dermatol; e-pub ahead of print 25 August 2010

www.jidonline.org 1985


Kyriakis KP, Michailides C, Palamara I et al. (2007) Lifetime prevalence distribution of chronic discoid lupus erythematosus. J Eur Acad Derm Venerol 21:1105–47 Lardet D, Martin S, Truchet F (2004) Effect of smoking on the effectiveness of antimalarial drugs for cutaneous lesions of patients with lupus: assessment in a prospective study. Rev Med Interne 25:786–91 Moghadam-Kia S, Chilek K, Gaines E et al. (2009) Cross-sectional analysis of a collaborative web-based database for lupus erythematosus-associated skin lesions: prospective enrollment of 114 patients. Arch Dermatol 145:255–60 Miot HA, Bartoli Miot LD, Haddad GR (2005) Association between discoid lupus erythematosus and smoking. Dermatology 211:118–22 Munster T, Gibbs JP, Shen D et al. (2002) Hydroxychloroquine concentrationresponse relationships in patients with rheumatoid arthritis. Arthritis Rheum 46:1460–9

main isoforms catalyzing N-desethylchloroquine formation. Drug Met Dispos 31:748–54 Rahman P, Gladman DD, Urowitz MB (1998) Smoking interferes with efficacy of antimalarial therapy in cutaneous lupus. J Rheumatol 25:1716–9 Tebbe B, Orfanos CE (1997) Epidemiology and socioeconomic impact of skin disease in lupus erythematosus. Lupus 6:96–104 Tett SE, Cutler DJ, Beck C et al. (2000) Concentration-effect relationship of hydroxychloroquine in patients with rheumatoid arthritis – a prospective, dose ranging study. J Rheumatol 27:1656–60 Tett SE, Day RO, Cutler DJ (1993) Concentration-effect relationship of hydroxychloroquine in rheumatoid arthritis – a cross-sectional study. J Rheumatol 20:1874–9

Niemi M, Leathart JB, Neuvorten A et al. (2003) Polymorphism in CYP2C8 is associated with reduced plasma concentrations of repaglinide. Clin Pharmacol Ther 74:380–7

Vera-Recabarren MA, Garcı´a-Carrasco M, Ramos-Casals M et al. (2010) Comparative analysis of subacute cutaneous lupus erythematosus and chronic cutaneous lupus erythematosus: clinical and immunological study of 270 patients. Br J Dermatol 162:91–101

Parikh S, Ouedraogo JB, Goldstein JA et al. (2007) Amodiaquine metabolism is impaired by common polymorphisms in CYP2C8: implications for malaria treatment in Africa. Clin Pharmacol Ther 82:197–203

Wahie S, McColl E, Reynolds NJ et al. (2010) Measuring disease activity and damage in discoid lupus erythematosus. Br J Dermatol 162:1030–7

Projean D, Baune B, Farinotti R et al. (2003) In vitro metabolism of chloroquine: identification of CYP 2C8, CYP 3A4, and CYP2D6 as the

Walling HW, Sontheimer RD (2009) Cutaneous lupus erythematosus: issues in diagnosis and treatment. Am J Clin Dermatol 10:365–81

1986 Journal of Investigative Dermatology (2011), Volume 131