Clinical and Laboratory Investigations
Dermatology 2005;211:118–122
DOI: 10.1159/000086440
Association between Discoid Lupus
erythematosus and Cigarette Smoking
A Case-Control Study
Hélio Amante Miota Luciane Donida Bartoli Miota
Gabriela Roncada Haddadb
a
Dermatology Department and b Internal Medicine, UNESP Medical School, Botucatu, Brazil
Key Words
Cigarette smoking ! Discoid lupus erythematosus
Abstract
Background: Discoid lupus erythematosus (DLE) is a
chronic cutaneous disease affecting photoexposed ar-
eas and has also been associated with cigarette smoking.
Objective: To evaluate the association between smoking
and DLE. Methods: A case-control study was performed
involving 57 cases diagnosed with DLE and 215 healthy
controls. Results: A higher smoking prevalence was not-
ed in DLE cases (84.2%) than controls (33.5%), and the
odds ratio adjusted for gender, age and ultraviolet index
in the city of origin was 14.4 (95% confidence interval
6.2–33.8; multiple logistic regression, p ! 0.01). The cu-
mulative smoking exposure was not related to prema-
ture DLE development. At the beginning of the disease,
smokers had more extensive involvement than non-
smokers; compromise of the upper arms was statisti-
cally related to smoking. Conclusion: Cigarette smoking
was statistically associated with DLE development. Oth-
er studies are needed in order to evaluate the effects of
smoking cessation on the course of disease.
Copyright © 2005 S. Karger AG, Basel
© 2005 S. Karger AG, Basel
1018–8665/05/2112–0118$22.00/0
Fax +41 61 306 12 34
E-Mail karger@karger.ch Accessible online at:
www.karger.com www.karger.com/drm
Received: July 15, 2004
Accepted: December 20, 2004
Introduction
Discoid lupus erythematosus (DLE) is a chronic, be-
nign inflammatory skin disease characterized by discrete-
ly infiltrated erythematous scaly plaques of varying sizes,
well-defined borders and atrophic centers generally in-
volving the face and photoexposed areas. The plaques
evolve with scarring and a tendency to hypochromia. Le-
sions can be restricted to the face or head or be diffuse
[1]. The mean age for DLE development is between 20
and 40 years, with a female-to-male ratio of 2–3:1 [2].
Some authors consider DLE a benign and localized
form of systemic lupus erythematosus (SLE), since dis-
coid lesions occur in 5–15% of patients with the systemic
disease, sometimes appearing before its overt manifesta-
tion [1, 3].
The etiology is autoimmune with genetic, hormonal
and environmental factors involved. Cigarette smoking
was suggested as an environmental trigger that induces
DLE in genetically predisposed individuals [3]. The influ-
ence of smoking on the immune system can be demon-
strated in other inflammatory diseases such as SLE,
Crohn’s disease, psoriasis vulgaris, Goodpasture’s syn-
drome, Graves’ disease, primary biliary cirrhosis, rheu-
matoid arthritis, pustulous palmoplantar psoriasis and
Siriraj Medical Library, Mahidol University
198.143.39.97 - 1/27/2016 4:02:15 PM
Dr. Hélio Amante Miot
Departamento de Dermatologia da UNESP, SN
Campus Rubião Jr
Downloaded by:
Botucatu, SP 18618–000 (Brazil)
Tel./Fax +55 21 14 3882 4922, E-Mail heliomiot@uol.com.br
thromboangiitis obliterans. Conversely, smoking is also
associated with decreased inflammatory activity in idio-
pathic ulcerative colitis [4, 5].
This study aims to evaluate the association between
cigarette smoking and DLE.

Methods

A case-control study was performed. Epidemiological informa-

tion on DLE and smoking habits was obtained with a standard
questionnaire. Cases with a clinical and histopathological diagnosis
of DLE who did not meet 4 or more SLE criteria were selected from
the connective diseases outpatient clinic of Hospital das Clínicas,
UNESP Medical School, between July 2002 and October 2003 [6].
Four controls were selected for each case from spouses, close rela- Fig. 1. Smoking prevalence according to age distribution (!2 p !
tives and household members; they were paired with cases accord- 0.05).
ing to sex, age and city of origin [7]. Cases were personally inter-
viewed after their consultation, and controls were individually con-
tacted by telephone.
The authors considered current smokers as those who were
smoking at the time of the interview and had been smoking at least Table 1. Comparisons between cases and controls
4 cigarettes/day for 4 years. The odds ratio (OR) was adjusted for
possible confounding factors: age, sex and ultraviolet index from DLE Controls p value
the city of origin using multiple logistic regression [7]. The popula-
tion-attributable risk was calculated using the adjusted OR, and n 57 215
smoking prevalence in the control group represented prevalence in Sex
the population [8]. Females 40 (70.2) 147 (68.4) >0.05 (!2)
Data were recorded and analyzed with Bioestat 2.0™, consider- Males 17 (29.8) 68 (31.6)
ing a significance of p ! 0.05 [9]. Median age 8 SD, years 40.0810.1 39.0816.4 >0.05 (t test)
UVO
5 54 (94.7) 198 (92.1) >0.05 (Yates)
6 3 (5.3) 17 (7.9)
Results
p values are unadjusted; UVO = ultraviolet index from the city
of origin; figures in parentheses are percentages.
Initially, 61 cases were selected, but 4 were excluded
for lack of histopathological examination, inconclusive
clinical diagnosis or diagnosis of SLE. The 215 control
cases were representative in comparison to cases for gen-
der, age and ultraviolet index from the city of origin Table 2. Epidemiological description of cases
(table 1).
There were more female cases (2.4:1), but there was p value
no difference in smoking rates between sexes; by the way, Median age at the beginning
among controls smoking was more prevalent in men (!2 of DLE 8 SD, years 31.089.6
p ! 0.01). Smoking prevalence was significantly higher for Females 31.089.2 >0.05 (t test)
adults in DLE groups at any age (fig. 1). Males 30.0810.6
The median age for disease development was 31 years, Smokers 31.088.6
>0.05 (t test)
Nonsmokers 25.0813.8
with no differences according to sex or smoking exposure Median age when they started
(table 2). Only 1 case had a close relative (sister) with smoking 8 SD, years 14.085.3
DLE; this is consistent with the literature on familial oc- Females 14.085.7
>0.05 (t test)
currence (!1.8%). Males 16.084.2
DLE affected the face most (91.2%), followed by the Proportion of smokers 48 (84.2%)
Females 34 (85.0%)
upper arms, neck and ears (table 3). Upper and lower arm Males 14 (82.4%)
>0.05 (Fisher)
occurrence was only found in smokers; the prevalence of
Siriraj Medical Library, Mahidol University
198.143.39.97 - 1/27/2016 4:02:15 PM

Discoid Lupus erythematosus and Dermatology 2005;211:118–122 119

Cigarette Smoking
Downloaded by:
smoking in patients with upper arm lesions was the only Table 3. Topographical distribution (%) of initial lesions and rela-
one to reach statistical significance. There was no statisti- tionship to smoking
cal association between the lesion topography and sex or
Smokers Non- Total p value
age (Fisher p 1 0.05). The number of body sites affected smokers
by the disease at the outset was significantly higher in
smokers (166.0%) (table 3). Face 91.7 88.9 91.2 >0.05 (Fisher)
Most DLE cases had started smoking at an early age Upper arms 43.8 – 36.8 <0.03 (Fisher)
Neck 25.0 22.2 24.6 >0.05 (Fisher)
(median 14 years) and presented an average consumption
Ears 22.9 11.1 21.1 >0.05 (Fisher)
of 29 cigarettes per day. There was no difference in smok- Scalp 16.7 11.1 15.8 >0.05 Fisher
ing prevalence between sexes (table 2). All DLE cases who Back 12.5 – 10.5 >0.05 (Fisher)
smoked had started smoking before developing the dis- Others 8.3 – 7.0 >0.05 (Fisher)
ease, with a mean interval of 17.4 8 9.1 years, but there Mean count of
affected regions 2.2 1.3 2.1 <0.05
was no relationship between the number of cigarettes
(Mann-Whitney)
smoked or accumulated smoking exposure until disease
development (Pearson p 1 0.05). p values are unadjusted.
There was a higher smoking prevalence in cases (84.2%)
compared to controls (33.5%) (table 4), and the calcu-
lated OR after adjustment for sex, age and ultravio-
let index from the city of origin was 14.4 (95% confi-
Table 4. Association between smoking and DLE
dence interval, 6.2–33.8, multiple logistic regression,
p ! 0.01).
DLE Controls Total
The population-attributable risk was estimated as
81.8% (95% confidence interval, 65.7–91.0%). Smokers 48 (84.2%) 72 (33.5%) 120
Nonsmokers 9 (15.8%) 143 (66.5%) 152
Total 57 215 272
Discussion
Unadjusted OR = 10.6 (95% confidence interval, 4.9–22.8),
p < 0.01. OR adjusted for sex, age and ultraviolet index from the
The higher smoking prevalence in DLE patients city of origin = 14.4 (95% confidence interval, 6.2–33.8), p < 0.01,
strongly suggests a relationship with disease develop- multiple logistic regression.
ment.
In another controlled study with fewer participants,
the smoking prevalence in DLE cases was 82% with an
OR of 12.2. The same study showed a higher cigarette use
(1.4 packs/day) in cases compared to controls (0.7 packs/ Solar radiation is classically involved in the disease’s
day). These values are consistent with our study [3]. The pathogenesis. Lesions can be induced experimentally by
median age for starting smoking was very precocious (14 UVB and UVA radiation, and visible light. Lesions wors-
years) with a high average of 29 cigarettes/day. It repre- en in more than half the patients during summer, but in
sents an important accumulated smoking exposure that winter it can also happen in up to 10% of cases [11].
could be a disease trigger in predisposed individuals, par- Pregnancy can improve DLE lesions during the second
ticularly considering that all cases with smoking habits and third trimester and worsen them during the first tri-
developed the disease after they had started to smoke. mester and in the postpartum period, but its impact is
The pathological immune response in DLE lesions less important than in SLE [12]. Anxiety and emotional
shows a typical Th1 profile, with the predominance of factors can also precipitate lesions and induce treatment
T lymphocyte infiltrate and inflammatory cytokines: resistance. Because smoking exposure is significantly
IFN-", IL-1 and TNF-# [10]. Discoid lesions can be spon- higher in DLE patients, it was considered a risk factor in
taneous or triggered by factors such as trauma, sunburn, skin lesion development. This was especially true for
psychological stress, infections, exposure to cold or preg- heavy smokers [3].
nancy. Exacerbations often occur with sun exposure and An association with HLA-B7, B8, Cw7, DR2, DR3
traumas. Occasionally, drugs like isoniazid, penicillin, and DQw1 has been described, along with an increased
griseofulvin and dapsone can precipitate DLE lesions [2]. risk of developing the disease in certain gene combina-
Siriraj Medical Library, Mahidol University
198.143.39.97 - 1/27/2016 4:02:15 PM

120 Dermatology 2005;211:118–122 Miot/Bartoli Miot/Haddad

Downloaded by:
tions, such as HLA-Cw7, DR3, DQw1 and HLA B7,
HLA-Cw7 and DR3 [13].
Familial cases are uncommon, accounting for less than
4%. A documented case of monozygotic twins with DLE
and polymorphic light eruption suggests that genetic fac-
tors and somatic mutations are also involved in the dis-
ease’s pathogenesis [14].
The age at DLE onset, male-to-female ratio, familial
occurrence and topographic distribution of lesions in this
study were in accordance with the literature. However,
this study has the first report of a significant compromise
in smokers’ upper arms [1]. Smoking is the most impor-
tant environmental risk in global public health. Reduc-
tions in smoking rates would have a direct favorable im-
pact in more than 50 diseases [5]. Smoker prevalence is
tending to reduce in populations, but even so global con-
sumption is increasing. The prevalence of smokers in the
USA was 37.4% in 1970; this had been reduced to 25%
in 1993 [3].
Smoking is also a risk factor associated with other in-
flammatory diseases such as SLE, with an OR of 2.3,
rheumatoid arthritis (OR of 2.0), thromboangiitis oblit-
erans (OR of 34.8), palmar pustulosis (OR of 7.2), Crohn’s
disease (OR of 3.1) and others [4, 5, 15, 16]. By the way,
a recent meta-analysis found a statistical association be-
tween SLE and smoking only for current smokers, but not
former smokers [17].
How smoking precisely affects lymphocyte inflamma-
tory activity is yet unknown. Some T lymphocytes with
suppressor profiles appear to have depressed activity
while an augmented immune activity stimulates prolif-
eration of peripheral T lymphocytes and B cells. Poly-
clonal B cell activation is one of the theories explaining
autoimmune activity in SLE [3, 18].
Cigarette smoke contains more than 100 organic sub-
stances, many of them known carcinogens and toxic
chemicals. Also, smoke has a genotoxic effect to DNA and
oxidative effects in the human body, interfering in the
metabolism of prostaglandins and the vascular reactivity
[18–20].
It is hard to estimate the real prevalence of smoking in
the Brazilian population. Smoking habits are influenced
by sex, education, age and cultural, familial and social
aspects. The choice of more than one control for each case
was aimed at increasing statistical power and lowering
sampling errors to reduce the chances of selection bias, as
smoking has so many external influences [7].
In 1989, 26% of Brazilian women and 40% of men over
15 years old smoked. In other case-control studies in Bra-
zil, the prevalence ranged from 25 to 34% [21, 22]. The
Discoid Lupus erythematosus and
Cigarette Smoking
comparison of smoking prevalence of cases with these
historical controls will identify a stronger association be-
tween DLE and smoking because controls in this study
had higher smoking rates than Brazilian surveys.
From case-control studies, we can also estimate the
percentage of the population-attributable risk. Regarding
the prevalence of smoking in the population and the role
of smoking in DLE pathogenesis, 81.8% of DLE cases
could have been influenced by smoking and might have
been prevented if exposure to smoking had not existed.
It was noted that 3 single patients who had quit smoking
showed clinical improvement in the lesions and 1 a wors-
ening when taking up smoking again. This fact, already
mentioned in the literature, supports the hypothesis of
the participation of smoking in the pathogenesis of DLE
[23].
This study did not evaluate the consumption of caf-
feine, alcohol, analgesics or hypnotics, which have a high-
er use among smokers and could also play a role in the
pathogenesis of DLE. The ultraviolet exposure dose from
professional exposure and outdoor activities could not be
isolated because of imprecise data from the participants
[9].
Another feature of this study was the lack of racial or
phototype adjustment, because it was difficult to be pre-
cise about this information from the telephone interviews
with controls. We were also unable to accurately estimate
the use of any medication at the beginning of the disease.
Nevertheless, the strong association between smoking
and DLE found in this study (OR of 14.4) reduces the
possibility of this being a chance result or confounding
interference; it suggests that smoking can represent an
independent risk factor for DLE. The discrepancy found
between the unadjusted and adjusted OR was due to a
weaker prevalence of smoking among women of the con-
trol group in comparison to men.
Some publications have highlighted an increased ther-
apy resistance in DLE, and especially a diminished re-
sponse to antimalarials in smokers, probably due to an
increase in drug excretion. It also seems to have a direct
negative influence on disease course, over and above the
pharmacological interference. Therefore, cessation of
smoking could positively interfere in the progression of
DLE, leading patients to less toxic and less expensive
treatments [9, 7, 24].
These findings can also reinforce an opportunity for
health education against smoking. Early exposure to to-
bacco, greater disease extension in smokers, diminished
therapeutic response and the high number of cigarettes
consumed are indirect evidence that smoking can influ-
Siriraj Medical Library, Mahidol University
198.143.39.97 - 1/27/2016 4:02:15 PM
Dermatology 2005;211:118–122 121
Downloaded by:
ence the course of DLE. This study detected a statistical
association between cigarette smoking and DLE. New
controlled studies are needed in order to measure how
giving up smoking would affect the therapeutic response
and its impact on the clinical course of the disease.
References
1 Duarte AA: Lupus eritematoso cutâneo. An
Bras Dermatol 2001;76:655–671.
2 Rowell NR, Goodfield MJD: The connective
tissue diseases; in Champion RH, Burton JL,
Burns DA, Breathnach SM (eds): Rook/Wilkin-
son/Ebling Textbook of Dermatology. Oxford,
Blackwell Scientific Publications, 1998, vol 3,
pp 1437–1575.
3 Gallego H, Crutchfield CE 3rd, Lewis EJ, Gal-
lego HJ: Report of an association between dis-
coid lupus erythematosus and smoking. Cutis
1999;63:231–234.
4 Rahman M, Chowdhury AS, Fukui T, Hira K,
Shimbo T: Association of thromboangiitis ob-
literans with cigarette and bidi smoking in
Bangladesh: A case-control study. Int J Epide-
miol 2000;29:266–270.
5 Smith JB, Fenske NA: Cutaneous manifesta-
tions and consequences of smoking. J Am Acad
Dermatol 1996;34:717–732.
6 Tan EM, Cohen AS, Fries JF, Masi AT, Mc-
Shane DJ, Rothfield NF, Schaller JG, Talal N,
Winchester RJ: The 1982 revised criteria for
the classification of systemic lupus erythema-
tosus. Arthritis Rheum 1982;25:1271.
7 Hennekens CH, Buring JE: Epidemiology in
Medicine, ed 1. Boston, Little, Brown & Co,
1987.
8 Bruzzi P, Green SB, Byar DP, Brinton LA,
Schairer C: Estimation of population attribut-
able risk for multiple factors using case-control
data. Am J Epidemiol 1985;122:904–914.
122 Dermatology 2005;211:118–122
Acknowledgements
The authors are thankful to the editor and referees for important
contributions to this paper.
9 Ayres M, Ayres M Jr, Ayres DL, dos Santos AS:
Bioestat: 2.0 Aplicações estatísticas nas áreas
das ciências biológicas e médicas. Belém, So-
ciedade Civil Mamirauá e MCT – CNPq,
2000.
10 Toro JR, Finlay D, Dou X, Zheng SC, LeBoit
PE, Connolly MK: Detection of type 1 cyto-
kines in discoid lupus erythematosus. Arch
Dermatol 2000;136:1497–1501.
11 Velthuis PJ, van Weelden H, van Wichen D,
Baart de la Faille H: Immunohistopathology of
light-induced skin lesions in lupus erythemato-
sus. Acta Derm Venereol 1990;70:93–98.
12 Shlepakov VM: Effect of gestation in the course
of chronic lupus erythematosus. Soviet Med
1969;32:111–116.
13 Knop J, Bonsmann G, Kind P, Doxiadis I,
Vogeler U, Doxiadis G, Goerz G, Grosse-Wil-
de H: Antigens of the major histocompatibility
complex in patients with chronic discoid lupus
erythematosus. Br J Dermatol 1990;122:723–
728.
14 Wojnarowska F: Simultaneous occurrence in
identical twins of discoid lupus erythematosus
and polymorphic light eruption. J R Soc Med
1983;76:791–792.
15 Achutti A, Menezes AMB: Epidemiologia do
tabagismo; in Achutti A (ed): Guia Nacional de
Prevenção e Tratamento do Tabagismo. Rio de
Janeiro, Vitrô Comunicação & Editora, 2001,
pp 9–27.
16 Nagata C, Fujita S, Iwata H, Kurosawa Y, Ko-
bayashi K, Kobayashi M, Motegi K, Omura T,
Yamamoto M, Nose T, et al: Systemic lupus
erythematosus: A case-control epidemiologic
study in Japan. Int J Dermatol 1995;34: 333–
337.
17 Costenbader KH, Kim DJ, Peerzada J, Lock-
man S, Nobles-Knight D, Petri M, Karlson
EW: Cigarette smoking and the risk of system-
ic lupus erythematosus: A meta-analysis. Ar-
thritis Rheum 2004;50:849–857.
18 Mills CM: Cigarette smoking, cutaneous im-
munity, and inflammatory response. Clin Der-
matol 1998;16:589–594.
19 Alberg AJ: The influence of cigarette smoking
on circulating concentrations of antioxidant
micronutrients. Toxicology 2002; 180: 121–
137.
20 Trüeb RM: Association between smoking and
hair loss: Another opportunity for health edu-
cation against smoking? Dermatology 2003;
206:189–191.
21 Menezes AMB, Horta BL, Oliveira ALB,
Kaufmann RAC, et al: Risco de câncer de pul-
mão, laringe e esôfago atribuível ao fumo. Rev
Saúde Pública 2002;36:129–134.
22 Ministério da Saúde, Instituto Nacional de Al-
imentação e Nutrição (INAN): PNSN: es-
tatísticas sobre hábitos de fumo no Brasil. Bra-
sília, INAN, 1989.
23 Jewell ML, McCauliffe DP: Patients with cu-
taneous lupus erythematosus who smoke are
less responsive to antimalarial treatment. J Am
Acad Dermatol 2000;42:983–987.
24 Rahman P, Gladman DD, Urowitz MB: Smok-
ing interferes with efficacy of antimalarial ther-
apy in cutaneous lupus. J Rheumatol 1998;25:
1716–1719.
Siriraj Medical Library, Mahidol University
198.143.39.97 - 1/27/2016 4:02:15 PM
Miot/Bartoli Miot/Haddad
Downloaded by: