Prospective case-control study of chlamydia, legionella and mycoplasma infections in patients with pityriasis rosea

ARTICLE

The cause of pityriasis rosea (PR) is unknown. A recent double-blind placebo-controlled trial [1] reported the benefit of erythromycin in modifying the course of PR. One of the postulated mechanisms for the action of this macrolide is the eradication of bacteria susceptible to erythromycin [1].

An association between PR and bacterial infections has been proposed but unproven in earlier reports [2-5], as summarised in Table 1. However, control subjects were not recruited for most of these studies. Acute and convalescent samples to document changes in antibody (Ab) titres were not available for all patients studied. Chlamydia spp. causes infections mainly of the respiratory and genital tracts of young persons with similar age distribution as PR. To our best knowledge, no study has been reported to evaluate the association between Chlamydia spp. and PR.

We report here a prospective case-control study of PR patients and paired age-and-sex-matched controls where serology against Chlamydia, Legionella and Mycoplasma spp. were all investigated. Acute and convalescent serum specimens were available from all study subjects.

Aim

The aim of this study was to investigate the association between PR and Chlamydia pneumoniae, C. trachomatis, Legionella longbeachae, L. micdadei, L. pneumophila, and Mycoplasma pneumoniae infections.

Materials and methods

Between 1 March 2000 and 31 August 2001 (18 months), all patients diagnosed to have PR in a primary care setting were invited to join the study. The diagnosis was based on clinical grounds by a physician with training and certifications in dermatology and paediatrics. Lesional biopsy for histopathology was arranged for all atypical cases.

Clotted blood was collected from each patient at initial presentation and a convalescent sample was collected four weeks later. For each patient with PR, the next patient of the same sex and comparable age (± two calendar years) requiring blood collection for non-dermatological disease who consented to participate in the study was recruited as a control. We obtained informed written consent from all study and control subjects, or from the parents or legal guardians for minors. The protocol of this study was approved by the Ethics Committee, Faculty of Medicine, University of Hong Kong (EC 1497-00).

The following tests were performed for each serum sample from study and control subjects: immunoglobulin (Ig) M and IgG against C. pneumoniae and C. trachomatis Groups B (B-E-D), C (C-J-H-I), and I (G-F-K) (by micro-immunofluorescence); IgM and IgG against L. longbeachae, L. micdadei, and L. pneumophila serotypes 1-14 (by indirect fluorescent Ab); and IgM, IgG and IgA against M. pneumoniae (by enzyme immunoassay). The serology tests on the acute and convalescent patient samples and the control serum were performed in parallel but results were read "blinded" to the clinical information. In addition a complete blood count was performed on each patient.

We defined a significant Ab rise as four-fold increase in Ab titres for a bacteria species or serotype in the acute and convalescent specimens for a patient. We used Fisher's Exact Probability Test, two-tailed, to analyse qualitative data (seroprevalence). We employed the Wilcoxon Match-Pairs Signed-Ranks Test to analyse changes in Ab titres collectively.

Results

A total of 26 patients were recruited, 13 being study subjects with PR, and 13 being paired-matched control subjects. The patients with PR were aged between eight and 46 years (mean: 26.8 years). Three were minors aged 8, 9, and 15. Four (30.8%) were males and nine (69.2%) were females. Three were English subjects aged 15, 29 and 46. The others were Chinese. HHV-6 and HHV-7 results of eight of these patients have been reported previously [6, 7]. Lesional biopsy was performed on one patient with atypical PR features, revealing focal spongiosis with perivascular lymphocytic infiltrates compatible with PR. Biopsy was not performed for the other 12 patients with typical PR. The 13 paired age-and-sex-matched control subjects ranged in age from nine to 47 years (mean: 27.7 years). All were Chinese subjects.

The serological results are summarised in Table 2. No patient and no control subject had evidence of acute infection by any of the bacteria investigated. Two patients (an eight-year-old boy and a 24-year-old lady) had four-fold rise in IgG titres against Chlamydia pneumoniae, both with titres of 1:8 in the acute specimen rising to 1:32 in the convalescent specimen. The failure to detect IgM for these two patients is highly suggestive of absence of active C. pneumoniae infection [8]. Moreover, these two patients had no symptom or sign of chest infection. The antibody rise might be related to an amnestic immune response only.

No patient had IgM positive against L. longbeachae. Five patients had IgG detected against L. longbeachae, with titres ranging from 1:64 to 1:256. There was neither an overall rise nor drop in IgG titres against L. longbeachae for any patient. No patient had IgM positive against L. micdadei. Three patients had IgG detectable against L. micdadei, with no significant rise in titres. The seroprevalence of study subjects for L. micdadei was insignificantly higher (P = 0.220) than control subjects.

A 24-year old lady with PR had IgG and IgM undetectable against all serotypes of L. pneumophila in the acute specimen. In the convalescent specimen, IgG was positive at low titres against serotypes 1, 3 and 7, while IgM was positive at low titre against serotype 6. Cross-reactivity could have explained these results.

A 30-year-old lady with PR had IgM, IgG and IgA detectable against M. pneumoniae. However, there was no significant rise in IgG titres, and titres of IgM and IgA were low. The patient had no symptom or sign of chest infection.

Discussion

All the bacteria investigated do not appear to play an aetiological role in our 13 patients with PR. The seroprevalence and IgG titres of the study patients for the bacteria investigated were insignificantly different from those of control subjects. The seroprevalence of L. micdadei and IgG titres to C. pneumoniae are of particular interest. Both are higher for study patients than for control subjects. In both cases, the difference might have become statistically significant if our study had recruited more study and control subjects. However, from the results of individual patients, there is no evidence that any of the patients seropositive for L. micdadei or demonstrating a high IgG titre for C. pneumoniae did have active infection.

We believe that more studies to confirm the benefit of erythromycin and other antibiotics in PR are warranted. Should the observed benefit of erythromycin [1] be genuine, it does not appear from results of our study and previous studies that the eradication of chlamydia, legionella or mycoplasma is a likely mechanism. The roles of other bacteria susceptible to erythromycin such as streptococci remain to be explored.

There is some evidence that immune dysfunction plays an aetiological role in PR. Apart from antibiotic effects, erythromycin also has anti-inflammatory and immunomodulatary effects [9]. We therefore believe that these effects might contribute towards its action in PR.

We believe that recruiting patients in a primary care setting is important in our present study. None of our patients required hospitalisation. These patients would have been missed if our study were hospital based. Some patients with PR may be referred to a dermatology out-patient clinic. However, the waiting period usually exceeds one to two months in the local setting, and the rash would have usually faded by the time these patients are seen. Moreover, the acute blood sample could not have been available. Valid serology interpretations could not have been drawn from results of convalescent samples only.

A limitation of this study is that other bacteria susceptible to erythromycin such as streptococci were not investigated. Further studies have to be conducted to evaluate roles of these bacteria in the pathogenesis of PR. Another limitation is that only serology was employed as the investigation tool in this study. Cross-reaction with other microorganisms might lead to difficulties in data interpretation. Should we have more resources, a study based on both serology and DNA detection might be more rewarding.

CONCLUSION

In a case-control study of 13 patients with PR and 13 paired age-and-sex-matched controls, we failed to identify an association of PR with C. pneumoniae, C. trachomatis, L. longbeachae, L. micdadei, L. pneumophila, and M. pneumoniae infections. We conclude that infections caused by these bacteria are unlikely to play a significant role in the pathogenesis of PR. We recommend further studies of PR to be along other directions including viruses, immune dysfunction and other bacteria susceptible to erythromycin..

Article accepted on 22/11/01

REFERENCES

1. Sharma PK, Yadav TP, Gautam RK, et al. Erythromycin in pityriasis rosea: a double-blind, placebo-controlled clinical trial. J Am Acad Dermatol 2000; 42: 241-4.

2. Hudson LD, Adelman S, Lewis. CW. Pityriasis rosea. Viral complement fixation studies. J Am Acad Dermatol 1981; 4: 544-6.

3. Bonafe JL, Icart J, Perpere M, et al. Histopathologic, ultrastructural, immunologic and virologic study of Gibert's pityriasis rosea. Ann Dermatol Venereol 1982; 109: 855-61.

4. Ishibashi A, Ueda I, Fujita K. Pityriasis rosea Gibert and Mycoplasma pneumoniae infection. J Dermatol 1985; 12: 97-100.

5. Gjenero-Margan I, Vidovic R, Drazenovic V. Pityriasis rosea Gibert: detection of Legionella micdadei antibodies in patients. Eur J Epidemiol 1995; 11: 459-62.

6. Chuh AAT, Chiu SSS, Peiris JSM. Human herpesvirus 6 and 7 DNA in peripheral blood leukocytes and plasma in patients with pityriasis rosea by polymerase chain reaction - a prospective case control study. Acta Derm Venereol 2001; 81: 289-90.

7. Chuh AA, Peiris JS. Three cases of pityriasis rosea in children with no evidence of human herpesvirus 7 infection. Pediatr Dermatol 2001; 18: 381-3.

8. Martin RE, Bates JH. Atypical pneumonia. Infect Dis Clin North Am 1991; 5: 585-601.

9. Labro MT. Anti-inflammatory activity of macrolides: a new therapeutic potential? J Antimicrob Chemother 1998; 41 (suppl. B): 37-46.