ARTICLE
Auteur(s) : Olivia Boccara1, Laurence
Valeyrie-Allanore2, Béatrice Crickx1, Vincent
Descamps1
1Department of Dermatology, Bichat Claude Bernard
Hospital, Assistance Publique-Hôpitaux de Paris, 46 rue Henri
Huchard, ParisFax (+33): 1 40257303
2Department of Dermatology, Henri Mondor Hospital,
Assistance Publique-Hôpitaux de Paris, Créteil
accepté le 5 Juillet 2006
Kano et al. recently reported an association between anticonvulsant
hypersensitivity syndrome and hypogammaglobulinemia [1]. We had
observed this association in our previously reported cases. In our
first reported case of DRESS associated with Human Herpes Virus 6
reactivation, there was a transient hypogammaglobulinemia (5.5 g/dL
at admission). In this case phenobarbital was the culprit drug. In
our first series of DRESS associated with human herpesvirus 6
infection we observed a hypogammaglobulinemia in 4 cases out of 7
[3]. All of these patients were previously treated with
carbamazepin.To further investigate this possible association of
hypogammaglobulinemia with DRESS, we retrospectively studied 39 new
consecutive cases of DRESS recruited in two departments of
dermatology and compared the prevalence of hypogammaglobulinemia
with a control group of 52 consecutive patients with
erythroderma/exfoliative dermatitis of non-drug cause for whom a
serum protein electrophoresis was available. In the group of DRESS
patients we compared the characteristics of the patients with and
without hypogammaglobulinemia. Our aim was to ascertain the
association of DRESS and hypogammaglobulinemia as compared to other
causes of erythroderma. The second aim of the study was to
investigate whether this association was limited to
anticonvulsants.
Patients and methods
Gammaglobulin levels were measured by a serum protein
electrophoresis. In 39 cases out of the 78 consecutive cases of
DRESS, a serum protein electrophoresis had been carried out during
the hospitalization.
Hypogammaglobulinemia was defined by a serum value of
gammaglobulins under the normal lower value given by the two
laboratories: 6 g/L (Henri-Mondor Hospital) and 6.6 g/L
(Bichat Claude Bernard Hospital). Hypoalbuminemia was defined by a
serum value of albumin under the normal lower value given by both
laboratories: 39 g/L. (Henri-Mondor Hospital) and 31 g/L
(Bichat Claude Bernard Hospital).
56 patients (27 controls and 29 DRESS) were included in
Henri-Mondor Hospital. 35 patients (25 controls and 10 DRESS) were
included in Bichat Claude Bernard Hospital.
We systematically investigated in these 39 DRESS patients (a)
the culprit drug, (b) associated diseases, (c) the course of the
DRESS, (d) sepsis. These 39 cases were compared to a control group
of 52 consecutive patients hospitalized during the same period with
erythroderma/exfoliative dermatitis with an available serum protein
electrophoresis. In the DRESS group, patients with
hypogammaglobulinemia (group 1) were compared to patients without
hypogammaglobulinemia (group 2).
The collected data were analyzed by using the Chi2
test to determine the p-value.
Results
In the 39 patients with DRESS, hypogammaglobulinemia was found in
12 patients (table 1)( Table 1 ). In
only three patients was a distant electrophoresis available which
showed that gammaglobulin levels returned to normal values. The
culprit drugs in these 12 cases (group 1) were mainly
anticonvulsant drugs: carbamazepin and phenytoin (table 2)( Table 2 ). But in 7 cases other causative
drugs were implicated: allopurinol, sulfasalazin, tazocillin,
ibuprofen, celecoxib, vancomycin.
Associated diseases which could participate in the
hypogammaglobulinemia in group 1 included: monoclonal gammapathy (2
patients) and rheumatoid arthritis (1 patient). In the patients of
group 2, some associated diseases might have contributed to the
absence of hypogammaglobulinemia: Human-Immunodeficiency Virus
infection (2 patients), Hepatitis B virus and Hepatitis C virus
infections (1 patient), malaria (1 patient). We never found
hypogammaglobulinemia when patients were from Asia (3 patients),
Africa or Caribbean Islands (11 patients) according to our
laboratory reference values. The mean age of the DRESS patients was
45.8 years.
In the control group of 52 patients with erythroderma (mean age
63.6 years) causes included: psoriasis (21 cases), atopic
dermatitis or eczema (9 cases), lymphoma or leukemia (8 cases),
actinoreticulosis (1 case), bullous dematosis (2 cases),
dermatomyositis (1 case), zinc deficiency (1 case), unknown
etiology (9 cases). Six patients had a hypogammaglobulinemia. Their
causes of erythroderma were leukemia (2 cases), psoriasis (2
cases), eczema (1 case) and lymphoma (1 case). The difference
between the patients with DRESS and the control group was
statistically significant for the presence of hypogammaglobulinemia
in DRESS (p = 0.022).
Mean serum albumin level was lower in patients with DRESS than
in patients with erythroderma (table 1) but hypogammaglobulinemia
was not correlated with hypoalbuminemia. In the DRESS group
hypoalbuminemia was observed in 7 patients out of the 12 with
hypogammaglobulinemia and in 11 patients out of the 27 patients
without hypogammaglobulinemia. In the same way in the non DRESS
group hypoalbuminemia was observed in 7 patients without
hypogammaglobulinemia (47 patients) and in 2 patients with
hypogammaglobulinemia (5 patients).
In patients with DRESS, the delay between the blood sampling for
the serum protein electrophoresis after drug withdrawal was longer
in patients without hypogammaglobulinemia as compared to patients
with hypogammaglobulinemia: 9.9 days versus 3.9 days (mean value),
respectively (table 2).
Among the 39 patients with DRESS, sepsis occurred in 5 patients:
3 of them had hypogammaglobulinemia. One of the two patients
without hypogammaglobulinemia had the blood sampling done very
late. The characteristics of the two groups of patient with DRESS
are described in table 2.
Table 1 Levels of serum gammaglobulins and albumin in
the two groups
|
DRESS
|
Erythroderma
|
|
Number of patients
|
39
|
52
|
|
Number of patients with hypogammaglobulinemia
|
12
|
6
|
|
Number of patients with hypoalbuminemia
|
18
|
9
|
|
Percentage of hypoalbuminemic patients with
hypogammaglobulinemia
|
39%
|
22%
|
|
Mean serum level of gammaglobulins (range)
|
11.6 g/L (3.5-48.3)
|
11.1 g/L (3.3-22.9)
|
|
Median serum level of gammaglobulins
|
9.1 g/L
|
10.2 g/L
|
|
Mean serum level of albumin
|
33.4 g/L (13.8-43.6)
|
35.6 g/L (23.2-49.7),
|
Table 2 Characteristics of the DRESS with or without
hypogammaglobulinemia
|
Group 1 Hypogammaglobulinemia
|
Group 2
|
|
Absence of hypogammaglobulinemia
|
|
Number of patients
|
12
|
27 (normogammaglobulinemia: 19, hypergammaglobulinemia: 8)
|
|
Number of patients with hypoalbuminemia
|
7
|
11
|
|
Ethnicity
|
African or Caribbean: 0
|
African or Caribbean: 11
|
|
Asian: 0
|
Asian: 3
|
|
Mean serum level of gammaglobulins
|
4.9 g/L
|
13.5 g/L
|
|
Mean serum level of albumin
|
31.7 g/L
|
32.7 g/L
|
|
Average delay of serum sampling after drug withdrawal (median)
|
3.9 days (4 days)
|
9.9 days (7.5 days)
|
|
Average delay between first drug intake and DRESS (median)
|
31 days (25 days)
|
32.4 days (28 days)
|
|
Average hospitalization duration (median)
|
8.7 days (9 days)
|
17 days; African and Caribbean: 22.6 days (10 days)
|
|
Carbamzepin
|
4
|
4
|
|
Phenytoin
|
1
|
0
|
|
Valproic acid
|
0
|
2
|
|
Lamotrigin
|
0
|
1
|
|
Minocycline
|
0
|
5
|
|
Allopurinol
|
1
|
8
|
|
Other drugs
|
6
|
7
|
Discussion
Hypogammaglobulinemia is frequent in DRESS and less frequent in
other causes of erythroderma. We demonstrated in this retrospective
study a statistically significant difference for the presence of
hypogammaglobulinemia in patients with DRESS as compared to other
causes of erythroderma (p = 0.022). Lower levels of serum
gammaglobulins and albumin were observed in DRESS patients as
compared to patients with erythroderma. But we consider that
hypogammaglobulinemia was not a simple consequence of
hypoproteinemia since hypogammaglobulinemia did not correlate with
hypoalbuminemia. Hypoproteinemia could participate in
hypogammaglobulinemia but did not explain by itself this
hypogammaglobulinemia. Prospective studies are necessary to confirm
the association of hypogammaglobulinemia with DRESS.
Moreover some cases of hypogammaglobulinemia could have been
missed in our patients with DRESS: (i) the blood sampling was later
in group 2 (DRESS patients without hypogammaglobulinemia at the
time of the blood sampling) as compared to group 1, (ii) normal
values of our laboratories may be not suitable for
African-Caribbean or Asian patients who may have higher rates of
gammaglobulins [4]. Some African patients with DRESS had higher
levels of gammaglobulins several weeks after the DRESS. These facts
may explain that the mean gammaglobulin levels were similar in
DRESS patients and controls (11.6 and 11.1 g/dl, respectively).
We do not think that these two groups of DRESS define two types
of DRESS. The course of the DRESS was similar in the 2 groups. The
delay between the first drug intake and DRESS was the same (table
2). The higher mean duration of hospitalization of patients in
group 2 was mainly explained by African-Caribbean patients who had
a more severe disease [5].
DRESS-associated hypogammaglobulinemia was not restricted to
anticonvulsant therapy: in our series various drugs were
potentially associated with hypogammaglobulinemia including
antibiotics (tazocillin and vancomycin), nonsteroidal
anti-inflammatory drugs (2 patients), and allopurinol.
This transient immune dysfunction is an interesting anomaly to
explain, as proposed by Kano et al., many characteristics of DRESS:
long-time duration for the development of DRESS after first drug
intake, spontaneous remission after drug withdrawal, relapse with
other drugs and reactivation of herpesvirus (Human herpesvirus 6,
Epstein-Barr virus, cytomegalovirus) [1]. But this
hypogammaglobulinemia could also be a risk factor for severe
bacterial infections. In our first reported case a severe
septicemia related to staphylococcal infection occurred within the
course of the DRESS [2].
However, DRESS pathogenesis is not clearly understood.
Hypogammaglobulinemia is probably a consequence of a severe B cell
depletion observed at the beginning of the DRESS induced by drugs
and other environmental factors (viruses) in a predisposed genetic
background [1, 2, 6, 7]. It is interesting to notice that most of
the DRESS-associated drugs are known to exhibit immunomodulatory
properties (minocycline, allopurinol, salazopyrin, anticonvulsants,
etc.). This association may have therapeutic consequences with the
use of intravenous gammaglobulins in DRESS, as proposed by Kano,
that warrants a prospective therapeutic study.
Acknowledgments
The authors sincerely thank JC Roujeau (Department of Dermatology,
Henri-Mondor Hospital) and E Begon for providing access and giving
the right to use the data from their patients.
References
1 Kano Y, Inaoka M, Shiohara T. Association between
anticonvulsant hypersensitivity syndrome and human herpesvirus 6
reactivation and hypogammaglobulinemia. Arch Dermatol 2004; 140:
183-8.
2 Descamps V, Bouscarat F, Laglenne S,
Aslangul E, Veber B, Descamps D, Saraux JL,
Grange MJ, Grossin M, Navratil E, Crickx B,
Belaich S. Human herpesvirus 6 infection associated with
anticonvulsant hypersensitivity syndrome and reactive
haemophagocytic syndrome. Br J Dermatol 1997; 137: 605-8.
3 Descamps V, Valance A, Edlinger C,
Fillet AM, Grossin M, Lebrun-Vignes B,
Belaich S, Crickx B. Association of human herpesvirus 6
infection with drug reaction with eosinophilia and systemic
symptoms. Arch Dermatol 2001; 137: 301-4.
4 Koh ET, Chi MS, Lowenstein FW. Comparison of
selected blood components by race, sex, and age. Am J Clin Nutr
1980; 33: 1828-35.
5 Begon E, Roujeau JC. Drug hypersensitivity syndrome:
DRESS (Drug Reaction with Eosinophilia and Systemic Symptoms. Ann
Dermatol Venereol 2004; 131: 293-7.
6 Aihara Y, Ito SI, Kobayashi Y, Yamakawa Y,
Aihara M, Yokota S. Carbamazepine-induced
hypersensitivity syndrome associated with transient
hypogammaglobulinaemia and reactivation of human herpesvirus 6
infection demonstrated by real-time quantitative polymerase chain
reaction. Br J Dermatol 2003; 149: 165-9.
7 Descamps V, Collot S, Mahe E, Houhou N,
Crickx B, Ranger-Rogez S. Active human herpesvirus 6
infection in a patient with drug rash with eosinophilia and
systemic symptoms. J Invest Dermatol 2003; 121: 215-6.
|
Printable version
Version PDF