Footprints
http://www.100md.com
《新英格兰医药杂志》
In this Journal feature, information about a real patient is presented in stages (boldface type) to an expert clinician, who responds to the information, sharing his or her reasoning with the reader (regular type). The authors' commentary follows.
A 42-year-old woman with a history of asthma and the irritable bowel syndrome presented with pain and swelling of her feet and ankles that were associated with a rash on her lower extremities. A few days earlier, the patient had awakened with an achy sensation in her feet, and the pain had progressed since then. She reported having difficulty in walking and noted she was "shuffling" to get from place to place. She had "painful red bumps" over her lower extremities. The rash, which was nonpruritic, started at the dorsum of the foot, extended over the malleoli bilaterally, and seemed to "march up the back" of her legs. Her feet were swollen and violaceous. She reported having a low-grade fever, migratory polyarthralgia, and intermittent loose stools that were nonbloody. During the previous week, she had been seen for periorbital swelling and nasal congestion and had started a 10-day course of trimethoprim–sulfamethoxazole for presumptive sinusitis.
The painful, bilateral nature of the patient's erythematous nodular rash on her lower extremities suggests erythema nodosum, a condition that has a broad differential diagnosis. The eye and nasal symptoms that had developed one week earlier may have been related to her current presentation. Erythema nodosum may be a manifestation of a hypersensitivity reaction to the trimethoprim–sulfamethoxazole that she was prescribed. Sulfa commonly causes drug reactions, and the one-week time frame is consistent with such reactions. Alternatively, the eye and nasal symptoms could be part of a multisystem rheumatologic condition that is now causing skin lesions and migratory polyarthralgia, such as sarcoidosis or Wegener's granulomatosis.
It would be important to review how her diagnosis of asthma was made, since dyspnea and cough could alternatively be symptoms of sarcoidosis. Sarcoidosis is frequently associated with erythema nodosum, particularly in young women. In addition, I would review her diagnosis of irritable bowel syndrome, since inflammatory bowel disease can be manifested as rash and arthritis. Erythema nodosum is the most common cutaneous manifestation of this disease. Finally, I would query the patient about other exposures, including medications. A complete blood count and differential count, with specific consideration of eosinophilia as a potential marker of hypersensitivity, would be an important starting point for the evaluation. I would also consider stool or serologic studies for pathogens associated with erythema nodosum (such as campylobacter, salmonella, yersinia, chlamydia, and mycoplasma).
The patient stopped taking trimethoprim–sulfamethoxazole because of concern about a drug reaction, and she was seen by a dermatologist. A biopsy was performed, and treatment with prednisone (40 mg daily, to be slowly tapered over a period of four weeks) was started for symptom relief. There was no cough, shortness of breath, pleurisy, abdominal pain, nausea, vomiting, night sweats, or weight loss. Her medical history was significant for asthma, which had been diagnosed clinically during childhood, and a presumptive diagnosis of the irritable bowel syndrome, which had been made several months earlier after she had presented with daytime loose stools that were nonbloody.
The patient's medications included albuterol and fluticasone inhalers, sertraline (for the premenstrual syndrome), and prednisone. She had no known drug allergies. She had been born in the United States and lived in a western Massachusetts suburb with her husband and two dogs. She had no children and had had two first-trimester spontaneous abortions. She did not smoke or use intravenous drugs, and she drank an average of four glasses of wine per week. Her travel history included trips to the Caribbean, to the Southwest, and to other regions of the northeastern United States. She reported that she had not had extensive exposure to soil or done any gardening. She and her husband had tested negative for the human immunodeficiency virus (HIV) and for syphilis before their marriage two years earlier. She had no family history of inflammatory bowel disease. However, her mother had rheumatoid arthritis.
Erythema nodosum is usually a clinical diagnosis and does not generally require biopsy unless atypical features are present. Perhaps the posterior distribution of the patient's rash prompted the biopsy. Therapy with corticosteroids is also generally not required, since the symptoms are self-limiting and the discomfort usually responds to nonsteroidal antiinflammatory drugs. Now that corticosteroids have been initiated, it is particularly important to ensure that underlying infection has been ruled out in order to avoid exacerbation of an untreated disease. The systemic infections that are associated with erythema nodosum are linked to specific geographic and environmental exposures. These include tuberculosis, coccidioidomycosis (which is found in the southwestern United States), and other endemic fungi, such as histoplasmosis (which is found in the Ohio and Mississippi river valleys). Streptococcal infection is commonly associated with erythema nodosum and may be the unifying factor linking the patient's rash and joint symptoms.
At this point, I would obtain a chest radiograph to evaluate the patient for an occult pulmonary infection or sarcoidosis. I would review the details of her eye symptoms for any history of photophobia, ophthalmalgia, conjunctival injection, blurred vision, excessive lacrimation, or other findings that could suggest uveitis, since this finding would make sarcoidosis much more likely. I would also further investigate her risk of exposure to tuberculosis or fungal infection. Given her travel history, I would consider testing for urinary histoplasma antigen and for serum antibodies to coccidioides, in addition to ordering a test for antistreptolysin O antibodies, liver-function tests, and a purified protein derivative (PPD) tuberculin test. The PPD test result may have limited value in a patient who is taking corticosteroids; however, a positive result would be very informative.
The patient was referred to a specialist in infectious diseases. Further occupational history was obtained, which revealed exposure to sealed blood samples at a same-day delivery service and the presence of a center for homeless people that was located near the patient's office, with shared lobby and elevator space. One month after the initiation of corticosteroid therapy, the patient reported resolution of her nodular rash and had tapered her dose of prednisone to 5 mg per day. The physical examination revealed a low-grade fever (37.2°C) and faint residual violaceous discoloration of the medial midfoot, with mild ankle tenderness bilaterally. The patient's chest was clear to auscultation, and the cardiac examination revealed no murmur. There were no joint effusions. Prednisone was discontinued.
Measurements of angiotensin-converting–enzyme and calcium levels, tests of renal and liver function, and foot radiographs were normal. A urinalysis was negative, as were tests for rheumatoid factor and antinuclear antibody. A complete blood count was unremarkable, with no evidence of eosinophilia (although the test was performed while the patient was taking corticosteroids). A PPD tuberculin test resulted in more than 20 mm of induration. A chest radiograph suggested subtle hilar fullness.
The patient continued to have mild symptoms without a definitive diagnosis. The size of the induration on the PPD test correlates with the specificity of the test for Mycobacterium tuberculosis; the larger the reaction, the more likely that the patient has been exposed to tuberculosis. The size of the reaction, however, does not indicate whether she has active or latent disease. The occupational history may be an important link to a workplace exposure to tuberculosis, and I would be interested in the results of prior PPD tests, as well as any history of exposure to an active case. The hilar fullness may also suggest sarcoidosis but, along with the positive PPD test, increases suspicion for tuberculosis.
Primary infection with tuberculosis can be manifested as mediastinal lymphadenopathy and is also associated with erythema nodosum. It is possible that the positive PPD test reflects only latent disease and that the hilar fullness seen on the chest radiograph is due to an alternative process; however, in view of the otherwise negative workup, tuberculosis must be ruled out. I would confirm hilar lymphadenopathy with a computed tomographic (CT) scan of the chest, since CT has greater sensitivity for detecting parenchymal or pleural involvement.
The skin biopsy showed features suggestive of erythema nodosum, but the presence of some lobular granulomatous inflammation (Figure 1A) and thickened vessel walls (Figure 1B) raised the question of erythema induratum. The findings were not typical of the latter, however, as there was a minimal inflammatory component and rare vessel involvement. Special stains for acid-fast bacilli and fungi were negative. A CT scan of the chest, abdomen, and pelvis showed nonspecific hilar and mediastinal lymphadenopathy (Figure 2). There was no evidence of parenchymal, upper-lobe, or cavitary disease and no intra-abdominopelvic abscesses, ascites, bony disease, or other signs of active tuberculosis. Stains of three induced sputum specimens were negative for acid-fast bacilli. Tests for urinary histoplasma antigen and for serum antibodies to coccidioides and a repeated HIV test were negative. The erythrocyte sedimentation rate was 5 mm per hour. The patient was afebrile and noted continued improvement in her lower-extremity symptoms even after corticosteroid treatment had been discontinued.
Figure 1. Skin-Biopsy Specimen from a Patient with Erythema Nodosum.
Panel A shows a focus of lobular granulomatous inflammation with a multinucleated giant cell (arrow). Acid-fast staining (not shown) revealed no mycobacteria. Panel B shows medium-size vessels with thickened walls (arrow).
Figure 2. CT Scan of the Chest.
The scan shows nonspecific hilar (Panel A, arrows) and mediastinal (Panel B, arrows) lymphadenopathy.
The skin-biopsy findings are consistent with the presence of erythema nodosum, with features suggestive of erythema induratum, although there is insufficient evidence of lobular inflammation and active vasculitis to provide strong support for the latter diagnosis. The distinction is relevant, since erythema induratum may be more specific for tuberculosis than is erythema nodosum; although controversial, this diagnosis, in association with other findings suggestive of tuberculosis, might support the initiation of treatment for possible active tuberculosis. The absence of acid-fast bacilli is not surprising; the mycobacteria typically are not present in tuberculosis-related cases of either erythema nodosum or erythema induratum. Although the association of erythema induratum with tuberculosis has been debated in the literature, there are case reports of polymerase-chain-reaction (PCR) assays that are positive for M. tuberculosis in biopsy specimens with histologic evidence of erythema induratum. I would not recommend ordering a PCR assay of tissue in this case, however, since a negative finding might be due to inadequate sensitivity of the test, and a positive finding might reflect amplification of mycobacterial fragments rather than viable mycobacterial organisms.
In summary, we have a patient with erythema nodosum, a low-grade fever, lymphadenopathy, and a positive PPD test. Given the many negative tests, tuberculosis is the most likely diagnosis. Erythema nodosum that is associated with tuberculosis typically occurs at the time of PPD conversion, and the patient's presentation is consistent with primary tuberculosis. In the majority of patients with primary tuberculosis, the disease evolves into a latent form. If this patient's symptoms persist, there are two options: treat the tuberculosis empirically and follow the clinical response of lymphadenopathy and skin disease or obtain lymph-node tissue for a definitive diagnosis. If her symptoms continue to abate, I would defer therapy and continue to follow her closely, with the expectation of a transition to latent tuberculosis. In the meantime, I would contact the homeless center located near her office and inquire about any recent active cases of tuberculosis among its clients.
The patient's lower-extremity symptoms completely resolved during the next few weeks. She did not have any further fever and reported having no chills, sweats, weight loss, or cough. Cultures of acid-fast bacilli from the three induced sputum tests were all negative. A few months later, while the patient was still asymptomatic and not taking any medications, a nine-month course of isoniazid was prescribed for latent tuberculosis. A follow-up CT scan of the chest obtained after the completion of treatment for latent tuberculosis was within normal limits. At one year, the patient was still not taking corticosteroids, and there had been no subsequent flares of migratory polyarthralgia, fever, or other systemic symptoms.
Commentary
The clinical triad of erythema nodosum, polyarthralgia, and hilar lymphadenopathy in a young woman with a history of asthma initially suggested L?fgren's syndrome associated with sarcoidosis. Erythema nodosum is associated with a variety of conditions, including sarcoidosis, rheumatologic disease, inflammatory bowel disease, an adverse reaction to medication, cancer, pregnancy, and infection. The discussant appropriately recognized several relevant etiologic considerations in this case, underscoring the importance of revisiting prior empirical clinical diagnoses (in this case, asthma, sinusitis, and the irritable bowel syndrome) in the context of new symptoms and findings. In addition to sarcoidosis, conditions that merited early consideration included sulfa-mediated hypersensitivity reaction, inflammatory bowel disease, bacterial enteritis, tuberculosis, Wegener's granulomatosis, parvovirus B19 infection, and endemic fungal infection. The differential diagnosis was effectively narrowed by consideration of geographic and epidemiologic factors in concert with laboratory and other test results.
Among panniculitides in women, erythema nodosum is the most common clinicopathological finding, followed by erythema induratum (or nodular vasculitis), Weber–Christian disease, Takayasu's arteritis, and other rheumatologic processes.1 Although the majority of cases of erythema nodosum are attributed to idiopathic causes, approximately one third of cases have an identifiable cause.2 Diagnostic assessment is an important part of the management of erythema nodosum, and rational evaluation includes a test for antistreptolysin O antibodies, a PPD test, and a chest radiograph.3
The initiation of corticosteroid treatment for symptomatic relief of the patient's painful rash and arthralgia proved effective in this case but could have been harmful. Nonsteroidal antiinflammatory drugs are generally adequate to mitigate the symptoms of erythema nodosum, without incurring the risk of uncontrolled infection. The positive PPD test was an important turning point in the evaluation, but the test was obtained only after the initiation of corticosteroids. The majority of cases of tuberculosis in the United States (including more than 75 percent of the cases in Massachusetts4) occur in foreign-born people, and tuberculosis was not initially suspected in this white woman of North American descent. A detailed occupational history was necessary to reveal the patient's risk of exposure.
Although not diagnostic, the skin-biopsy findings suggesting erythema induratum raised further suspicion of a tuberculosis-related process. In retrospect, the posterior distribution of lesions over the patient's calves may have been an early clue, since this site is commonly involved in erythema induratum. Other features of erythema induratum include a tendency to ulcerate and to recur, but neither of these features was present in this case. Specific findings of lobular inflammation and active vasculitis are necessary to make this diagnosis. Erythema nodosum, by contrast, is a clinical diagnosis that rarely requires biopsy. Although classically described as occurring on the shins, the rash can occur in any distribution.
The nature of the association of tuberculosis with erythema induratum is controversial.5,6 The amplification of M. tuberculosis DNA from cutaneous tissue has had variable results.7,8 Stains and cultures of acid-fast bacilli are often negative in both erythema induratum and erythema nodosum, as seen in this case. Abnormal findings on chest radiographs and an elevated erythrocyte sedimentation rate have been reported in the majority of cases of erythema induratum and erythema nodosum that are associated with active tuberculosis.9 When erythema nodosum occurs, it generally does so at the time of PPD conversion and often precedes other symptoms of tuberculosis. The pathophysiology is thought to reflect a delayed-type hypersensitivity reaction associated with primary disease. In a small percentage of patients, active tuberculosis develops, but most patients will have latent infection.
The key to this case was determining the nature of the relationship between the positive PPD test and the skin findings. Was the erythema nodosum a coincidental finding (i.e., related to trimethoprim–sulfamethoxazole therapy and independent of the latent tuberculosis), a reactive feature (i.e., a delayed-type hypersensitivity reaction at the time of PPD conversion), or a feature of active tuberculosis involving the skin? This distinction was critical in order to determine the appropriate treatment. Hilar lymphadenopathy and epidemiologically plausible exposure made tuberculosis-related erythema nodosum the most likely alternative. Initially, one could not distinguish between active, latent, or primary tuberculosis. Only further evaluation and follow-up could clarify the status of the infection. Pointing to a diagnosis of latent disease were the negative serial sputum tests for acid-fast bacilli, the absence of pulmonary parenchymal involvement and active extrapulmonary disease, the resolution of fever, the absence of other systemic symptoms over time, and the fact that the disease did not progress in the absence of antituberculosis therapy while the patient was taking corticosteroids.
Lymph-node biopsy might have yielded a more definitive diagnosis earlier in the course of the disease and simultaneously assessed for the possibility of sarcoidosis, but it was reasonable to forgo this invasive procedure in light of the complete resolution of the patient's symptoms. The clinicians initiated treatment for latent tuberculosis only after observing the patient over time and conducting a thorough search for active disease and other infections or inflammatory conditions that may be associated with erythema nodosum. Because erythema nodosum can precede other signs or symptoms of tuberculosis, it was prudent to delay treatment until the clinician had determined whether the patient had active or latent tuberculosis. There is no urgency to treat latent tuberculosis, and the risk of treating active tuberculosis with a single agent is drug resistance. Nine months of isoniazid therapy and close clinical follow-up were important features of the patient's care.
Acquisition of tuberculosis is generally asymptomatic, and PPD conversions are typically detected retrospectively during routine screening. However, in a small subgroup of patients, acute symptoms may develop, including erythema nodosum, fever, and lymphadenopathy, at the time of acquisition of tuberculosis infection. Although it is unusual to identify such patients at the time of PPD conversion, erythema nodosum provided an early "footprint" of tuberculosis in this case. The finding of erythema nodosum should prompt a focused search for possible triggers, particularly in cases in which corticosteroid therapy is considered. Attention to such footprints may help uncover atypical cases of infectious or inflammatory illnesses in unexpected hosts.
Source Information
From the Division of Infectious Diseases, Beth Israel Deaconess Medical Center and Harvard Medical School (S.K.); the Department of Immunology and Infectious Diseases, Harvard School of Public Health (J.P.D.); and the Division of Infectious Diseases, Brigham and Women's Hospital and Harvard Medical School (J.P.D.) — all in Boston.
Address reprint requests to Dr. Kassutto at the Division of Infectious Diseases, Beth Israel Deaconess Medical Center, Boston, MA 02215, or at skassutto@bidmc.harvard.edu.
References
Handa R, Ramam M, Verma KK, et al. Panniculitides -- a clinicopathologic study. J Assoc Physicians India 2002;50:1008-1012.
Cribier B, Caille A, Heid E, Grosshans E. Erythema nodosum and associated diseases: a study of 129 cases. Int J Dermatol 1998;37:667-672.
Garcia-Porrua CG, Gonzalez-Gay MA, Vasquez-Caruncho M, et al. Erythema nodosum: etiologic and predictive factors in a defined population. Arthritis Rheum 2000;43:584-592.
Trends in tuberculosis morbidity-United States, 1992-2002. MMWR Morb Mortal Wkly Rep 2003;52:217-20, 222.
Cribier B, Gorosshans E. Bazin's erythema induratum: obsolete concept and terminology. Ann Dermatol Venereol 1990;117:937-943.
Cho KH, Lee DY, Kim YG. Erythema induratum of Bazin. Int J Dermatol 1996;35:802-808.
Tan SH, Tan BH, Goh CL, et al. Detection of Mycobacterium tuberculosis DNA using polymerase chain reaction in cutaneous tuberculosis and tuberculids. Int J Dermatol 1999;38:122-127.
Tan SH, Tan HH, Sun YJ, Goh CL. Clinical utility of polymerase chain reaction in the detection of Mycobacterium tuberculosis in different types of cutaneous tuberculosis and tuberculids. Ann Acad Med Singapore 2001;30:3-10.
Boonchai W, Suthipinittharm P, Mahaisavariya P. Panniculitis in tuberculosis: a clinicopathologic study of nodular panniculitis associated with tuberculosis. Int J Dermatol 1998;37:361-363.
Related Letters:
Footprints
Al-Khasawneh K., White P. Jr., Kassutto S., Daily J. P.(Sigall Kassutto, M.D., an)
A 42-year-old woman with a history of asthma and the irritable bowel syndrome presented with pain and swelling of her feet and ankles that were associated with a rash on her lower extremities. A few days earlier, the patient had awakened with an achy sensation in her feet, and the pain had progressed since then. She reported having difficulty in walking and noted she was "shuffling" to get from place to place. She had "painful red bumps" over her lower extremities. The rash, which was nonpruritic, started at the dorsum of the foot, extended over the malleoli bilaterally, and seemed to "march up the back" of her legs. Her feet were swollen and violaceous. She reported having a low-grade fever, migratory polyarthralgia, and intermittent loose stools that were nonbloody. During the previous week, she had been seen for periorbital swelling and nasal congestion and had started a 10-day course of trimethoprim–sulfamethoxazole for presumptive sinusitis.
The painful, bilateral nature of the patient's erythematous nodular rash on her lower extremities suggests erythema nodosum, a condition that has a broad differential diagnosis. The eye and nasal symptoms that had developed one week earlier may have been related to her current presentation. Erythema nodosum may be a manifestation of a hypersensitivity reaction to the trimethoprim–sulfamethoxazole that she was prescribed. Sulfa commonly causes drug reactions, and the one-week time frame is consistent with such reactions. Alternatively, the eye and nasal symptoms could be part of a multisystem rheumatologic condition that is now causing skin lesions and migratory polyarthralgia, such as sarcoidosis or Wegener's granulomatosis.
It would be important to review how her diagnosis of asthma was made, since dyspnea and cough could alternatively be symptoms of sarcoidosis. Sarcoidosis is frequently associated with erythema nodosum, particularly in young women. In addition, I would review her diagnosis of irritable bowel syndrome, since inflammatory bowel disease can be manifested as rash and arthritis. Erythema nodosum is the most common cutaneous manifestation of this disease. Finally, I would query the patient about other exposures, including medications. A complete blood count and differential count, with specific consideration of eosinophilia as a potential marker of hypersensitivity, would be an important starting point for the evaluation. I would also consider stool or serologic studies for pathogens associated with erythema nodosum (such as campylobacter, salmonella, yersinia, chlamydia, and mycoplasma).
The patient stopped taking trimethoprim–sulfamethoxazole because of concern about a drug reaction, and she was seen by a dermatologist. A biopsy was performed, and treatment with prednisone (40 mg daily, to be slowly tapered over a period of four weeks) was started for symptom relief. There was no cough, shortness of breath, pleurisy, abdominal pain, nausea, vomiting, night sweats, or weight loss. Her medical history was significant for asthma, which had been diagnosed clinically during childhood, and a presumptive diagnosis of the irritable bowel syndrome, which had been made several months earlier after she had presented with daytime loose stools that were nonbloody.
The patient's medications included albuterol and fluticasone inhalers, sertraline (for the premenstrual syndrome), and prednisone. She had no known drug allergies. She had been born in the United States and lived in a western Massachusetts suburb with her husband and two dogs. She had no children and had had two first-trimester spontaneous abortions. She did not smoke or use intravenous drugs, and she drank an average of four glasses of wine per week. Her travel history included trips to the Caribbean, to the Southwest, and to other regions of the northeastern United States. She reported that she had not had extensive exposure to soil or done any gardening. She and her husband had tested negative for the human immunodeficiency virus (HIV) and for syphilis before their marriage two years earlier. She had no family history of inflammatory bowel disease. However, her mother had rheumatoid arthritis.
Erythema nodosum is usually a clinical diagnosis and does not generally require biopsy unless atypical features are present. Perhaps the posterior distribution of the patient's rash prompted the biopsy. Therapy with corticosteroids is also generally not required, since the symptoms are self-limiting and the discomfort usually responds to nonsteroidal antiinflammatory drugs. Now that corticosteroids have been initiated, it is particularly important to ensure that underlying infection has been ruled out in order to avoid exacerbation of an untreated disease. The systemic infections that are associated with erythema nodosum are linked to specific geographic and environmental exposures. These include tuberculosis, coccidioidomycosis (which is found in the southwestern United States), and other endemic fungi, such as histoplasmosis (which is found in the Ohio and Mississippi river valleys). Streptococcal infection is commonly associated with erythema nodosum and may be the unifying factor linking the patient's rash and joint symptoms.
At this point, I would obtain a chest radiograph to evaluate the patient for an occult pulmonary infection or sarcoidosis. I would review the details of her eye symptoms for any history of photophobia, ophthalmalgia, conjunctival injection, blurred vision, excessive lacrimation, or other findings that could suggest uveitis, since this finding would make sarcoidosis much more likely. I would also further investigate her risk of exposure to tuberculosis or fungal infection. Given her travel history, I would consider testing for urinary histoplasma antigen and for serum antibodies to coccidioides, in addition to ordering a test for antistreptolysin O antibodies, liver-function tests, and a purified protein derivative (PPD) tuberculin test. The PPD test result may have limited value in a patient who is taking corticosteroids; however, a positive result would be very informative.
The patient was referred to a specialist in infectious diseases. Further occupational history was obtained, which revealed exposure to sealed blood samples at a same-day delivery service and the presence of a center for homeless people that was located near the patient's office, with shared lobby and elevator space. One month after the initiation of corticosteroid therapy, the patient reported resolution of her nodular rash and had tapered her dose of prednisone to 5 mg per day. The physical examination revealed a low-grade fever (37.2°C) and faint residual violaceous discoloration of the medial midfoot, with mild ankle tenderness bilaterally. The patient's chest was clear to auscultation, and the cardiac examination revealed no murmur. There were no joint effusions. Prednisone was discontinued.
Measurements of angiotensin-converting–enzyme and calcium levels, tests of renal and liver function, and foot radiographs were normal. A urinalysis was negative, as were tests for rheumatoid factor and antinuclear antibody. A complete blood count was unremarkable, with no evidence of eosinophilia (although the test was performed while the patient was taking corticosteroids). A PPD tuberculin test resulted in more than 20 mm of induration. A chest radiograph suggested subtle hilar fullness.
The patient continued to have mild symptoms without a definitive diagnosis. The size of the induration on the PPD test correlates with the specificity of the test for Mycobacterium tuberculosis; the larger the reaction, the more likely that the patient has been exposed to tuberculosis. The size of the reaction, however, does not indicate whether she has active or latent disease. The occupational history may be an important link to a workplace exposure to tuberculosis, and I would be interested in the results of prior PPD tests, as well as any history of exposure to an active case. The hilar fullness may also suggest sarcoidosis but, along with the positive PPD test, increases suspicion for tuberculosis.
Primary infection with tuberculosis can be manifested as mediastinal lymphadenopathy and is also associated with erythema nodosum. It is possible that the positive PPD test reflects only latent disease and that the hilar fullness seen on the chest radiograph is due to an alternative process; however, in view of the otherwise negative workup, tuberculosis must be ruled out. I would confirm hilar lymphadenopathy with a computed tomographic (CT) scan of the chest, since CT has greater sensitivity for detecting parenchymal or pleural involvement.
The skin biopsy showed features suggestive of erythema nodosum, but the presence of some lobular granulomatous inflammation (Figure 1A) and thickened vessel walls (Figure 1B) raised the question of erythema induratum. The findings were not typical of the latter, however, as there was a minimal inflammatory component and rare vessel involvement. Special stains for acid-fast bacilli and fungi were negative. A CT scan of the chest, abdomen, and pelvis showed nonspecific hilar and mediastinal lymphadenopathy (Figure 2). There was no evidence of parenchymal, upper-lobe, or cavitary disease and no intra-abdominopelvic abscesses, ascites, bony disease, or other signs of active tuberculosis. Stains of three induced sputum specimens were negative for acid-fast bacilli. Tests for urinary histoplasma antigen and for serum antibodies to coccidioides and a repeated HIV test were negative. The erythrocyte sedimentation rate was 5 mm per hour. The patient was afebrile and noted continued improvement in her lower-extremity symptoms even after corticosteroid treatment had been discontinued.
Figure 1. Skin-Biopsy Specimen from a Patient with Erythema Nodosum.
Panel A shows a focus of lobular granulomatous inflammation with a multinucleated giant cell (arrow). Acid-fast staining (not shown) revealed no mycobacteria. Panel B shows medium-size vessels with thickened walls (arrow).
Figure 2. CT Scan of the Chest.
The scan shows nonspecific hilar (Panel A, arrows) and mediastinal (Panel B, arrows) lymphadenopathy.
The skin-biopsy findings are consistent with the presence of erythema nodosum, with features suggestive of erythema induratum, although there is insufficient evidence of lobular inflammation and active vasculitis to provide strong support for the latter diagnosis. The distinction is relevant, since erythema induratum may be more specific for tuberculosis than is erythema nodosum; although controversial, this diagnosis, in association with other findings suggestive of tuberculosis, might support the initiation of treatment for possible active tuberculosis. The absence of acid-fast bacilli is not surprising; the mycobacteria typically are not present in tuberculosis-related cases of either erythema nodosum or erythema induratum. Although the association of erythema induratum with tuberculosis has been debated in the literature, there are case reports of polymerase-chain-reaction (PCR) assays that are positive for M. tuberculosis in biopsy specimens with histologic evidence of erythema induratum. I would not recommend ordering a PCR assay of tissue in this case, however, since a negative finding might be due to inadequate sensitivity of the test, and a positive finding might reflect amplification of mycobacterial fragments rather than viable mycobacterial organisms.
In summary, we have a patient with erythema nodosum, a low-grade fever, lymphadenopathy, and a positive PPD test. Given the many negative tests, tuberculosis is the most likely diagnosis. Erythema nodosum that is associated with tuberculosis typically occurs at the time of PPD conversion, and the patient's presentation is consistent with primary tuberculosis. In the majority of patients with primary tuberculosis, the disease evolves into a latent form. If this patient's symptoms persist, there are two options: treat the tuberculosis empirically and follow the clinical response of lymphadenopathy and skin disease or obtain lymph-node tissue for a definitive diagnosis. If her symptoms continue to abate, I would defer therapy and continue to follow her closely, with the expectation of a transition to latent tuberculosis. In the meantime, I would contact the homeless center located near her office and inquire about any recent active cases of tuberculosis among its clients.
The patient's lower-extremity symptoms completely resolved during the next few weeks. She did not have any further fever and reported having no chills, sweats, weight loss, or cough. Cultures of acid-fast bacilli from the three induced sputum tests were all negative. A few months later, while the patient was still asymptomatic and not taking any medications, a nine-month course of isoniazid was prescribed for latent tuberculosis. A follow-up CT scan of the chest obtained after the completion of treatment for latent tuberculosis was within normal limits. At one year, the patient was still not taking corticosteroids, and there had been no subsequent flares of migratory polyarthralgia, fever, or other systemic symptoms.
Commentary
The clinical triad of erythema nodosum, polyarthralgia, and hilar lymphadenopathy in a young woman with a history of asthma initially suggested L?fgren's syndrome associated with sarcoidosis. Erythema nodosum is associated with a variety of conditions, including sarcoidosis, rheumatologic disease, inflammatory bowel disease, an adverse reaction to medication, cancer, pregnancy, and infection. The discussant appropriately recognized several relevant etiologic considerations in this case, underscoring the importance of revisiting prior empirical clinical diagnoses (in this case, asthma, sinusitis, and the irritable bowel syndrome) in the context of new symptoms and findings. In addition to sarcoidosis, conditions that merited early consideration included sulfa-mediated hypersensitivity reaction, inflammatory bowel disease, bacterial enteritis, tuberculosis, Wegener's granulomatosis, parvovirus B19 infection, and endemic fungal infection. The differential diagnosis was effectively narrowed by consideration of geographic and epidemiologic factors in concert with laboratory and other test results.
Among panniculitides in women, erythema nodosum is the most common clinicopathological finding, followed by erythema induratum (or nodular vasculitis), Weber–Christian disease, Takayasu's arteritis, and other rheumatologic processes.1 Although the majority of cases of erythema nodosum are attributed to idiopathic causes, approximately one third of cases have an identifiable cause.2 Diagnostic assessment is an important part of the management of erythema nodosum, and rational evaluation includes a test for antistreptolysin O antibodies, a PPD test, and a chest radiograph.3
The initiation of corticosteroid treatment for symptomatic relief of the patient's painful rash and arthralgia proved effective in this case but could have been harmful. Nonsteroidal antiinflammatory drugs are generally adequate to mitigate the symptoms of erythema nodosum, without incurring the risk of uncontrolled infection. The positive PPD test was an important turning point in the evaluation, but the test was obtained only after the initiation of corticosteroids. The majority of cases of tuberculosis in the United States (including more than 75 percent of the cases in Massachusetts4) occur in foreign-born people, and tuberculosis was not initially suspected in this white woman of North American descent. A detailed occupational history was necessary to reveal the patient's risk of exposure.
Although not diagnostic, the skin-biopsy findings suggesting erythema induratum raised further suspicion of a tuberculosis-related process. In retrospect, the posterior distribution of lesions over the patient's calves may have been an early clue, since this site is commonly involved in erythema induratum. Other features of erythema induratum include a tendency to ulcerate and to recur, but neither of these features was present in this case. Specific findings of lobular inflammation and active vasculitis are necessary to make this diagnosis. Erythema nodosum, by contrast, is a clinical diagnosis that rarely requires biopsy. Although classically described as occurring on the shins, the rash can occur in any distribution.
The nature of the association of tuberculosis with erythema induratum is controversial.5,6 The amplification of M. tuberculosis DNA from cutaneous tissue has had variable results.7,8 Stains and cultures of acid-fast bacilli are often negative in both erythema induratum and erythema nodosum, as seen in this case. Abnormal findings on chest radiographs and an elevated erythrocyte sedimentation rate have been reported in the majority of cases of erythema induratum and erythema nodosum that are associated with active tuberculosis.9 When erythema nodosum occurs, it generally does so at the time of PPD conversion and often precedes other symptoms of tuberculosis. The pathophysiology is thought to reflect a delayed-type hypersensitivity reaction associated with primary disease. In a small percentage of patients, active tuberculosis develops, but most patients will have latent infection.
The key to this case was determining the nature of the relationship between the positive PPD test and the skin findings. Was the erythema nodosum a coincidental finding (i.e., related to trimethoprim–sulfamethoxazole therapy and independent of the latent tuberculosis), a reactive feature (i.e., a delayed-type hypersensitivity reaction at the time of PPD conversion), or a feature of active tuberculosis involving the skin? This distinction was critical in order to determine the appropriate treatment. Hilar lymphadenopathy and epidemiologically plausible exposure made tuberculosis-related erythema nodosum the most likely alternative. Initially, one could not distinguish between active, latent, or primary tuberculosis. Only further evaluation and follow-up could clarify the status of the infection. Pointing to a diagnosis of latent disease were the negative serial sputum tests for acid-fast bacilli, the absence of pulmonary parenchymal involvement and active extrapulmonary disease, the resolution of fever, the absence of other systemic symptoms over time, and the fact that the disease did not progress in the absence of antituberculosis therapy while the patient was taking corticosteroids.
Lymph-node biopsy might have yielded a more definitive diagnosis earlier in the course of the disease and simultaneously assessed for the possibility of sarcoidosis, but it was reasonable to forgo this invasive procedure in light of the complete resolution of the patient's symptoms. The clinicians initiated treatment for latent tuberculosis only after observing the patient over time and conducting a thorough search for active disease and other infections or inflammatory conditions that may be associated with erythema nodosum. Because erythema nodosum can precede other signs or symptoms of tuberculosis, it was prudent to delay treatment until the clinician had determined whether the patient had active or latent tuberculosis. There is no urgency to treat latent tuberculosis, and the risk of treating active tuberculosis with a single agent is drug resistance. Nine months of isoniazid therapy and close clinical follow-up were important features of the patient's care.
Acquisition of tuberculosis is generally asymptomatic, and PPD conversions are typically detected retrospectively during routine screening. However, in a small subgroup of patients, acute symptoms may develop, including erythema nodosum, fever, and lymphadenopathy, at the time of acquisition of tuberculosis infection. Although it is unusual to identify such patients at the time of PPD conversion, erythema nodosum provided an early "footprint" of tuberculosis in this case. The finding of erythema nodosum should prompt a focused search for possible triggers, particularly in cases in which corticosteroid therapy is considered. Attention to such footprints may help uncover atypical cases of infectious or inflammatory illnesses in unexpected hosts.
Source Information
From the Division of Infectious Diseases, Beth Israel Deaconess Medical Center and Harvard Medical School (S.K.); the Department of Immunology and Infectious Diseases, Harvard School of Public Health (J.P.D.); and the Division of Infectious Diseases, Brigham and Women's Hospital and Harvard Medical School (J.P.D.) — all in Boston.
Address reprint requests to Dr. Kassutto at the Division of Infectious Diseases, Beth Israel Deaconess Medical Center, Boston, MA 02215, or at skassutto@bidmc.harvard.edu.
References
Handa R, Ramam M, Verma KK, et al. Panniculitides -- a clinicopathologic study. J Assoc Physicians India 2002;50:1008-1012.
Cribier B, Caille A, Heid E, Grosshans E. Erythema nodosum and associated diseases: a study of 129 cases. Int J Dermatol 1998;37:667-672.
Garcia-Porrua CG, Gonzalez-Gay MA, Vasquez-Caruncho M, et al. Erythema nodosum: etiologic and predictive factors in a defined population. Arthritis Rheum 2000;43:584-592.
Trends in tuberculosis morbidity-United States, 1992-2002. MMWR Morb Mortal Wkly Rep 2003;52:217-20, 222.
Cribier B, Gorosshans E. Bazin's erythema induratum: obsolete concept and terminology. Ann Dermatol Venereol 1990;117:937-943.
Cho KH, Lee DY, Kim YG. Erythema induratum of Bazin. Int J Dermatol 1996;35:802-808.
Tan SH, Tan BH, Goh CL, et al. Detection of Mycobacterium tuberculosis DNA using polymerase chain reaction in cutaneous tuberculosis and tuberculids. Int J Dermatol 1999;38:122-127.
Tan SH, Tan HH, Sun YJ, Goh CL. Clinical utility of polymerase chain reaction in the detection of Mycobacterium tuberculosis in different types of cutaneous tuberculosis and tuberculids. Ann Acad Med Singapore 2001;30:3-10.
Boonchai W, Suthipinittharm P, Mahaisavariya P. Panniculitis in tuberculosis: a clinicopathologic study of nodular panniculitis associated with tuberculosis. Int J Dermatol 1998;37:361-363.
Related Letters:
Footprints
Al-Khasawneh K., White P. Jr., Kassutto S., Daily J. P.(Sigall Kassutto, M.D., an)