To the Editor: In a recent issue of the Journal, Chew et al described a woman from Vietnam with skin nodules that, on histological examination, showed lobular panniculitis with granulomatous inflammation.1 No mycobacteria were visible and a polymerase chain reaction test for Mycobacterium tuberculosis was negative. Two months after starting quadruple antituberculous therapy (including rifampicin), her lesions had resolved. Erythema induratum (ostensibly due to hypersensitivity to M. tuberculosis) was diagnosed, despite the absence of evidence of tuberculosis. Other possible diagnoses were considered, but leprosy was not mentioned.

In regions of Australia where leprosy is not endemic, the disease is frequently overlooked.2 Birrell3 described a man from Malta with recurring skin lumps. Biopsy showed panniculitis with giant cells, and the man was initially misdiagnosed as having “Weber–Christian syndrome” or “relapsing febrile non-suppurative nodular panniculitis”. Soon after, another Maltese patient presented similarly. This time, leprosy was suggested, and a biopsy revealed the presence of Mycobacterium leprae.4 Re-examination of slides from the first case showed similar organisms, confirming leprosy.5

The patients described by Chew et al and Birrell had migrated from countries in which leprosy was endemic, and biopsies revealed granulomatous panniculitis. Weber–Christian syndrome and erythema induratum are rare, ill-defined conditions with confused aetiologies, and both lack a specific diagnostic test. Therefore, cases of leprosy can be easily misdiagnosed as one of these conditions. That the biopsy in this patient did not show visible M. leprae is against a diagnosis of leprosy. But in my experience, even in lepromatous (multibacillary) disease, occasionally a skin smear of a lesion or (more rarely) a biopsy specimen may fail to reveal bacilli. Of course, this would be likely if the patient had received specific treatment for leprosy previously.

Respectfully, I suggest that Chew et al should attempt to exclude lepromatous leprosy in their patient by looking for possible missed stigmata of leprosy, enquiring whether she has ever been treated for leprosy, asking whether any close acquaintances have had the infection or a chronic skin condition, and following up the patient in the long term.

To the Editor: One of the Journal’s recent Lessons from Practice illustrates common errors in the approach to dermatological conditions.1 As in all areas of medicine, an accurate diagnosis is crucial to the management of any skin disease. This is especially so if a medical practitioner institutes treatments, such as oral steroids, that have considerable potential for causing morbidity. The lessons I would draw from the case of erythema induratum described are as follows.

If you suspect an unusual presentation of a common condition, perform investigations to confirm your suspicions. Although erythema nodosum classically occurs on the anterior lower leg, lesions above the knee may occasionally be seen.

To make a diagnosis, investigations need to be appropriate. The battery of blood tests ordered in the case described would not have shed light on the pathological process occurring in the skin. There is a reluctance among the general medical community to perform skin biopsies. These procedures cause little morbidity, have a high diagnostic yield, and should be within the skill set of any medical graduate. Concern over causing a scar is often cited as a reason for not doing a biopsy. But, in my experience, patients are rarely worried about such a prospect. Missing the diagnosis is surely of much greater concern. Taking a simple biopsy, including fat, at the initial presentation would have saved the patient in question a lot of trouble and risk.

If there is no response to your treatment, it may well be that the initial diagnosis was incorrect. For example, it is common to see “steroid-resistant eczema” that is actually intraepidermal carcinoma. Erythema nodosum will usually show at least some response to non-steroidal anti-inflammatory treatment. Lack of response to a treatment that usually works should lead to a re-evaluation of the diagnosis.

Systemic steroids should not be used for a dermatological condition without a firm diagnosis. Firstly, they can suppress many of the clinical and histological changes that allow a diagnosis to be made. Appropriate investigations need to be done before starting steroids. Secondly, a drug like prednisolone may well make matters worse, especially if, as here, there is an infectious aetiology.

Patients from areas in which tuberculosis is endemic should have this condition excluded before being given systemic steroids. A lack of obvious exposure to or symptoms of tuberculosis is not unusual in patients from such areas who are subsequently shown to harbour this infection.

The authors state that, as erythema induratum can resolve with corticosteroid treatment, this can lead to an erroneous diagnosis of erythema nodosum. Using response to treatment as a quasi-diagnostic test is dangerous indeed. Steroids will cause many conditions associated with significant inflammation to improve or even appear to resolve. But this does not mean that there is no infectious or malignant aetiology.

In reply: We thank Bennett and Muir for their pertinent comments. Our patient did not have any history or clinical evidence of lepromatous leprosy. The skin biopsy did not reveal any dermal granulomatous involvement, and there were definitely no organisms seen on an auramine stain of the biopsy specimen.

Subcutaneous involvement in leprosy is uncommon except in erythema nodosum leprosum or as a neurotropic phenomenon. When present, it tends to be a neutrophil-rich hypersensitivity necrotising vasculitis — no features of which were seen in this case. Neither the woman’s partner nor child had a chronic skin condition or clinical history of leprosy or tuberculosis. Furthermore, the patient has been followed up for 12 months, with no recurrence of the rash.

We agree with Muir that an accurate diagnosis is crucial to managing any skin disease and that there were many lessons to be gathered from this case apart from the five points we listed. It is our usual practice not to begin definitive treatment until we have examined a skin biopsy of any suspicious lesion and made a diagnosis. As this patient was very concerned about getting a scar, we did not perform a skin biopsy initially, but informed her that we may need to do so if the condition did not respond to treatment.

We agree that patients from areas where tuberculosis is endemic should have tuberculosis excluded before instituting systemic steroid treatment. In this case, the patient was given a chest x-ray by the appropriate authorities before her migration to Australia. She has not returned to Vietnam since then. Furthermore, the patient had failed a trial of a non-steroidal anti-inflammatory drug and found the lesions cosmetically distressing. Consequently corticosteroids were instituted.