COMPARISON OF ORAL ERYTHROMYCIN, LOCAL ADMINISTRATION OF STREPTOMYCIN AND PLACEBO THERAPY FOR NONSUPPURATIVE BACILLUS CALMETTE-GUÉRIN LYMPHADENITIS

Abstract

The most common complications of Bacillus Calmette-Guérin (BCG) vaccination are regional lymphadenitis and abscess formation at the site of vaccination.1 The incidence of BCG-induced lymphadenitis is reported to be 0.5 to 5% of the vaccinated population, and it is significantly more common in newborns than in older children.1, 2 The necessity and mode of treatment for this complication, however, remain controversial. Treatment modalities include no treatment, surgical drainage or excision, needle aspiration and systemic antituberculous drugs or erythromycin.3-7 Recently local isoniazid instillation was reported to be superior to oral erythromycin therapy and placebo in the treatment of suppurative BCG lymphadenitis.8 However, a parenteral form of isoniazid is not currently available in many countries. On the other hand local use of streptomycin, an easily accessible drug, has not been previously evaluated. The aim of this randomized, controlled trial was to evaluate the role of local streptomycin instillation in treating this complication. Patients and methods. The study group comprised 45 patients (23 male, 22 female) who were seen at the Dr. Sami Ulus Children's Hospital between June, 1996, and January, 1997, with nonsuppurative BCG lymphadenitis. Dr. Sami Ulus Children's Hospital is a large reference hospital with annual admissions of ∼90 000 patients ages 0 to 15 years from not only Ankara but all over the country. The diagnostic criterion for BCG lymphadenitis was an ipsilateral axillary or supraclavicular lymphadenitis with a diameter of at least 1.5 cm and no other identifiable cause other than BCG vaccination which preceded the adenitis by no more than 6 months. The criteria required for inclusion were: (1) age between 2 months and 6 months; (2) having BCG lymphadenitis without signs of suppurative changes (increase in the size of the lymph node, redness in the skin and induration) or fluctuation; (3) no health problems; and (4) no treatment for lymphadenitis before referral. Forty-five patients fulfilling these criteria were enrolled into the study after we obtained written consent from the parents and ethical committee of the hospital. All patients underwent a complete physical examination, including measuring node size and BCG scar size. Node size was determined by measuring the largest transverse diameter of the lymphadenopathy. Chest radiographs and tuberculin skin tests were performed. The entire study population had been vaccinated for BCG at birth or 2 months of age with 0.05 ml of Danish strain 1331 (containing a minimum of 2 × 106 colony-forming units/ml bacilli) intradermally at the distal third of the deltoid region. This strain has not been evaluated for susceptibility to erythromycin in our country. The patients were assigned in a random order to each of three study groups: Group A patients (n = 15) were followed without any therapy and served as controls; Group B patients (n = 15) received oral erythromycin stearate (40 mg/kg/day in four divided doses) for 1 month; Group C patients (n = 15) received one dose (20 mg/kg) of streptomycin directly into the node. Needle aspiration was performed when fluctuation of the node was noted during follow-up of the patients in Groups B and C. Suppuration of the lymph node was determined with increase in the size of the lymph node, redness in the skin and induration. To avoid iatrogenic sinus formation, the local instillation and aspiration procedures were done subcutaneously 2 to 3 cm from the periphery of the node. The aspirated specimens were examined by Gram-stained smear and Ziehl-Neelsen stain and were plated for regular bacterial and mycobacterial growth using standard techniques. Mycobacterial species identification was performed by growth on Lowenstein-Jensen medium. All patients were examined every 2 weeks for determination of the node size, changes in the node consistency and detection of any complications related to therapy. At each evaluation the volume of medication consumed by each subject was checked by the author to assess compliance in taking erythromycin. Healing was defined as the regression of the lymph node to <0.5 cm without evidence of residual ulceration and sinus formation and no clinical or radiographic changes attributable to tuberculosis. In data analysis the values of the different groups were compared, two at a time, by the unpaired Student t test. The alpha value used was 0.05. Results. Clinical and laboratory findings relevant to each group of patients are presented in Table 1. The physical examinations were normal except for lymphadenitis, and chest radiographs were normal. The numbers of patients vaccinated for BCG either at birth or at the end of the second month were 14 and 1 in the Group A, 13 and 2 in Group B and 12 and 3 in Group C, respectively. When the groups were considered together the lymphadenitis appeared at the age of 3.15 ± 0.2 (mean ± SD) months for the infants vaccinated at birth, whereas that age was 4.50 ± 0.2 (mean ± SD) months for the infants vaccinated after 2 months of age (P = 0.019).TABLE 1: Clinical and laboratory characteristics of patient groups* The localization of the lymphadenopathy was left axillary region in 39, supraclavicular region in 6 and both regions in 6 patients. In Group A 3 patients (20%) showed resolution of the lymphadenitis without drainage. However, abscess formation with spontaneous perforation developed in the remaining 12 patients (80%), 6 (40%) of whom eventually showed sinus formation and ulceration. All the patients in Groups B and C showed progression to abscess formation and fluctuation of the lymph node, but fistulation or ulceration was not encountered. Lymph node aspiration was performed in the entire population of Groups B and C when the lymph node was fluctuant. Specimens obtained by aspiration did not have positive cultures. Gram-stained smears of the aspirates from 5 patients of Group B and 2 patients of Group C demonstrated leukocytes in abundant numbers. Ziehl-Neelsen stain of the same specimens was positive for acid-fast bacilli in 1 (6.66%) patient of Group B and 2 (13.3%) patients of Group C. Times to development of various responses according to treatment groups are presented in Table 2. Both Group B and Group C patients developed suppuration of the node significantly earlier than Group A. There were no statistical differences between the average time to suppuration of Group B and that of Group C and also the average time to aspiration (fluctuation) of the same groups (P = 0.341) after initiation of therapy. The mean time to healing among Group B patients was similar to mean that of Group A. However, Group C patients showed significantly shorter healing times than those of Groups B and A.TABLE 2: Time of various responses after therapy No side effects of local streptomycin treatment were observed during this study. Discussion. Infection with BCG spreads from the inoculation site via the lymphatic system to local lymph nodes and produces immunity against natural primary infection, with different protection rates for various forms of tuberculosis.9 Some authors think that unless BCG induces lymphadenitis to some extent no protective immune response has been induced.10 Development of BCG-induced lymphadenitis can be influenced by factors such as the type of vaccine, dose, concentration, storage duration of vaccine and route of administration.11, 12 Even if not treated, the natural course of BCG-induced lymphadenitis, whether suppurative or not, is eventual regression in most cases. Without therapy these lesions may persist for several weeks or months and cause discomfort to the child.13 Approximately 40 to 70% of nontreated cases of BCG adenitis, especially if appearing within the first 2 months, spontaneously perforate and drain. This protracted drainage eventually results in ulcerative lesions in a considerable number of cases.3, 4, 6, 7 In our study 80% of cases followed without therapy showed spontaneous drainage and 40% developed sinus tract formation and ulceration. Different modes of treatment for BCG lymphadenitis have been suggested and tried.3-8, 13, 14 Medical therapy, especially erythromycin, although it may have some beneficial effect in those cases developing lymphadenitis shortly after vaccination, is not recommended routinely.4, 7 Some authors advocate total excision of the node if the adenitis has developed within 2 months after vaccination or if the node diameter is >3 cm. Others propose that it should be reserved for those patients in whom other less aggressive measures have been unsuccessful.3-5, 7 Needle aspiration is useful in suppurative BCG lymphadenitis. This therapeutic intervention has not only prevented spontaneous drainage but also provided more rapid regression of the node.6, 14 The significantly high rate of complications4, 6, 7 and the rather long time (∼28 weeks) it takes for the node to regress spontaneously3, 6 point to the need to treat BCG adenitis cases in the nonsuppurative phase. In our study both of the treatment groups developed suppuration of the node in a shorter time than did the nontreated group. The mean duration until healing among the patients treated with oral erythromycin plus aspiration and those followed without any therapy were similar. However, cases managed with local streptomycin instillation plus needle aspiration showed significantly earlier resolution than the other two groups. Streptomycin instillation is also superior to oral erythromycin therapy in terms of patient compliance, because taking a drug four times in a day for a long period is more troublesome to the patient than a sole local injection. In conclusion local streptomycin instillation during the nonsuppurative phase of BCG adenitis plus aspiration is safe and effective treatment causing significantly earlier healing than oral erythromycin plus aspiration therapy and no treatment. Necdet Kuyucu, M.D. Semanur Kuyucu, M.D. Burhan Öcal, M.D. Tahsin Teziç, M.D. Dr. Sami Ulus Children's Hospital; Ankara, Turkey