Authors M. S. Jawahar (corresponding author), K. Rajaram, C. N. Paramasivan and K. Chandrasekar, Tuberculosis Research Centre, Mayor V R Ramanathan Road, Chetput, Chennai 600 031, India. Tel.: +91 44 2836 9611; E-mail: email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org S. Sivasubramanian, 123 Kamarajar Salai, Ramakrishna Nagar, Alwarthirunagar, Chennai 600 087, India. Tel.: +91 44 2486 3242; E-mail: email@example.com M. N. Kamaludeen, Department of Surgery, Government Rajaji Hospital, Madurai 625020, India. Tel.: +91 45 2252 9515; E-mail: firstname.lastname@example.org A. J. Thirithuvathas, 23 Pattimedu Road, Visalakshipuram, Madurai 625020, India. Tel.: 98430 51059; E-mail: email@example.com V. Ananthalakshmi, 16 Bharathi Park Road II, Avinashalingam University Post, Coimbatore 641 043, India. Tel.: +91 42 2245 0452 R. Prabhakar, K III, 16th Street, Block K3, L2 Apartments, Bogainvilla, Annanagar, Chennai 600 040, India. Tel.: +91 44 2626 4781
Objective The currently recommended treatment for lymph node tuberculosis is 6 months of rifampicin and isoniazid plus pyrazinamide for the first 2 months, given either daily or thrice weekly. The objective of this study was to assess the efficacy of a 6-month twice-weekly regimen and a daily two-drug regimen.
Methods Patients with biopsy confirmed superficial lymph node tuberculosis were randomly allocated to receive either a daily self-administered 6-month regimen of rifampicin and isoniazid, or a twice-weekly, directly observed, 6-month regimen of rifampicin and isoniazid plus pyrazinamide for the first 2 months, in Madurai, South India, Patients were followed up for 36 months after completing treatment.
Results Of 277 enrolled patients, data was available for analysis in 268. At the end of treatment, 116 of 134 [87%; 95% confidence interval (CI) 81–93%] patients in each treatment group had a favourable clinical response; 14 (11%; 95% CI 6–16%) and 17 (13%; 95% CI 7–19%) patients had a doubtful response, and 4 (3%; 95% CI 0–6%) and 1 (1%; 95% CI 0–2%) patients had an unfavourable response among those treated with the daily and twice-weekly regimen, respectively. During 36 months after completion of treatment, five patients [2 (2%; 95% CI 1–3%) and 3 (2%; 95% CI 1–3%) patients treated with the daily and twice-weekly regimen, respectively] had relapse of lymph node tuberculosis, of 260 assessed. Adverse reactions probably attributable to the treatment regimens occurred in 1% of the patients treated daily and in 11% of those treated twice-weekly (P < 0.001). At the end of 36 months after treatment, 126 of 134 (94%; 95% CI 90–98%) and 129 of 134 (96%; 95% CI 94–98%) of the patients treated with the daily and twice-weekly regimen, respectively, had a successful outcome.
Conclusion Both the self-administered daily regimen and the fully observed twice-weekly regimen were highly efficacious for treating patients with lymph node tuberculosis and may be considered as alternative options to the recommended regimens.
Lymphadenitis is the commonest extra-pulmonary manifestation of tuberculosis. In a series of randomized clinical trials, the British Thoracic Society (BTS) showed that the duration of treatment for this disease could be shortened from 18 months to 9 months initially and later to 6 months (Campbell & Dyson 1977,1979; British Thoracic Society Research Committee 1985,1988,1992; Campbell et al. 1993). A clinical trial by the Tuberculosis Research Centre, Chennai had previously shown that childhood tuberculous lymphadenitis could be successfully treated with a 6-month regimen that included streptomycin (Jawahar et al. 1990). The World Health Organisation (WHO) currently recommends a 6-month regimen consisting of rifampicin and isoniazid, plus pyrazinamide for the first 2 months, given daily or thrice-weekly for the treatment of patients with category-III tuberculosis which includes lymph node tuberculosis (World Health Organization 1997).
Twice-weekly 6-month regimens are effective for treating pulmonary tuberculosis (Rani 1991). A randomized clinical trial has demonstrated the efficacy of a daily 6-month regimen of rifampicin and isoniazid for treating spinal tuberculosis (Indian Council of Medical Research/British Medical Research Council Working Party in South India 1989). We hypothesized that similar regimens, i.e., a three-drug twice-weekly regimen or a two-drug daily regimen may be equally effective in lymph node tuberculosis since it is a paucibacillary disease. Such regimens will offer a great practical advantage, both for patients and health care providers.
Aim of study
We conducted a randomized clinical trial to study the efficacy of two 6-month regimens for treating patients with biopsy-confirmed, superficial lymph node tuberculosis: one a twice-weekly regimen of rifampicin and isoniazid, supplemented by pyrazinamide for the first 2 months, given under direct observation, and the other a daily self-administered regimen of rifampicin and isoniazid with once-weekly clinic attendance.
Materials and methods
Patients attending the paediatric and adult surgical out-patient departments of a large tertiary care hospital in Madurai, South India, with superficial lymphadenitis, were assessed for the trial. A lymph node was surgically excised from each patient and divided into two parts, one for histological examination and the other for mycobacterial culture using multiple culture media (Mitchison et al. 1983; Vanajakumar et al. 1997). Positive cultures were tested to identify the species of the mycobacteria by standard procedures (Kubica 1973) and susceptibility to isoniazid, streptomycin and rifampicin (Canetti et al. 1969). Patients with either histological or bacteriological evidence of tuberculosis were included in the trial if they (or their parent or guardian, in case of children), gave informed consent, were willing to attend the clinic for observed treatment and permitted home visits by the clinic staff. Patients with smear-positive pulmonary tuberculosis, or those who had received previous treatment for tuberculosis for longer than a month and those with evidence of renal or hepatic disease were not admitted to the study.
On enrollment, each participant had a full clinical examination, tuberculin test with 1 TU PPD RT23 and an X-ray of the chest (PA view). Two sputum specimens were examined by smear and culture for tubercle bacilli if patients had an abnormal chest X-ray.
Clinical examination was repeated every month up to 12 months, every 3 months from then until 24 months, and every 6 months thereafter till 36 months. Chest X-rays were taken at 1 month after starting treatment, at end of treatment, and at 12, 24 and 36 months, for those with initial abnormal films.
Using sealed envelopes containing numbered slips generated by random numbers, a statistician randomly assigned participants to the two study arms. Participants who received isoniazid and rifampicin daily for 6 months (6RH) collected the drugs once weekly for self-administration. The other group received isoniazid and rifampicin for 6 months plus pyrazinamide for the first 2 months, twice-weekly (2RHZ2/4RH2) under direct observation.
Adult patients (≥13 years of age) received isoniazid 300 mg (daily) or 600 mg (twice-weekly); rifampicin 450 mg (daily or twice-weekly) and pyrazinamide 1500 mg twice-weekly. Children (1–12 years of age) received a weight-adjusted dose, based on 5 mg/kg for isoniazid (daily) or 10 mg/kg (twice-weekly); 10 mg/kg for rifampicin (daily or twice-weekly) and 30 mg/kg for pyrazinamide (twice-weekly). Patients who defaulted for treatment were visited by a clinic staff or a letter was posted. Adherence to treatment among patients receiving the daily regimen was monitored by surprise home visits and pill counts.
An independent assessor, who was blinded to the regimen to which the patient had been allocated, examined all patients at the end of treatment, or earlier if considered necessary. He classified the clinical response to treatment as ‘favourable’, if there was disappearance or regression of lymph nodes to ≤10 mm diameter, and healing of sinuses and abscesses, if any; as ‘unfavourable’ if there was clinical deterioration or lack of clinical response after 4 months of treatment, or development of tuberculosis at other sites; or as ‘doubtful’ when lymph nodes regressed in size but were still palpable and exceeding 10 mm diameter with no other symptoms. Patients with doubtful response were advised repeat lymph node biopsies and were reviewed with the results of the biopsies. If culture of the node yielded Mycobacterium tuberculosis, the clinical response was revised as unfavourable and the patient was re-treated for tuberculosis. Patients whose lymph node histology was suggestive of tuberculosis but whose culture was negative, were kept under observation.
In patients with a favourable or doubtful response at end of treatment, ‘relapse’ was defined as recurrence of lymph node enlargement, either with biopsy-confirmed tuberculosis with histological or bacteriological proof of diagnosis, or in association with clinical deterioration. ‘Adverse reactions to treatment’ were those symptoms possibly attributable to the drugs used in the treatment regimens.
Sample size calculation
Sample size was calculated in consultation with the statistician. Assuming a favourable outcome in previous studies of lymph node TB in 95%, and expecting the cure rate of the test regimens to be 90% (since these are less intensive regimens), with a confidence limit factor of 95%, and applying the formula n = z2pq/l2, the number of patients required for each regimen was estimated to be 138. For a two-regimen study the required sample size was thus 278.
The proportions of patients with different outcome measures in the two regimens were compared using the chi-square test of significance with Yate's correction. Fisher's exact test was used whenever an expected value was less than five. A P-value ≤0.05 was considered statistically significant. An approximate 95% confidence interval (CI) was calculated using the formula P ± 1.96√pq/n.
The study protocol was approved by the institutional Scientific Advisory and Ethics Committees. Informed consent was obtained from all patients or their parents/guardians in the case of children.
Between June 1988 and September 1993, 831 patients who had lymph node biopsy suggestive of tuberculosis at the Madurai Rajaji Hospital were assessed for eligibility for the trial. Of these, 351 were unwilling to participate in the trial and 160 were from locations beyond the study site. Forty-three patients were found ineligible for reasons such as previous TB treatment, poor clinical condition, concomitant TB at other sites etc.
Of 277 patients enrolled to the study, nine were excluded (lost for treatment, seven; wrong diagnosis, one; non-TB death, one) and 268 were available for analysis of results at the end of treatment. During the 36 months after completion of treatment, two patients died from non-TB causes and one was lost for follow-up (Figure 1). Baseline patient characteristics are shown in Table 1. The median age was 10 years (range 1–55) for males and 20 years (range 1–65) for females.
Table 1. Baseline characteristics of 268 patients with lymph node tuberculosis admitted to a randomized clinical trial in Madurai, South India
Total patients n = 268
Daily n = 134
Twice-weekly n = 134
† Tuberculin reaction results were not available for nine patients.
‡ Lymph node culture results were not available for three patients.
Age and sex
Children (n = 87)
13–29 years (n = 144)
30–49 years (n = 35)
≥50 years (n = 4)
Tuberculin reaction to 1 TU PPD†
Lymph node culture for mycobacteria‡
Positive for M. tuberculosis
Drug susceptibility profile of M. tuberculosis cultures (n = 178)
Susceptible to isoniazid, streptomycin and rifampicin
Isoniazid and streptomycin
Isoniazid and rifampicin
Isoniazid and streptomycin and rifampicin
Sixty-three (72%) children and 171 (95%) adults had involvement of the cervical nodes, either alone or in association with other nodes (P < 0.001). Involvement of the cervical nodes alone was seen in 26 (30%) children and 136 (75%) adults (P < 0.001). Multiple sites (≥3 groups) were involved in 12 (14%) children and 11 (6%) adults. Lymph node abscesses or sinuses were present in 17 (20%) children and 14 (8%) adults (P < 0.01).
Fourteen patients, whose lymph node culture yielded non-tuberculous mycobacteria (NTM), were included in the analysis because the lymph node histology was suggestive of tuberculosis, and the culture of NTM was considered to be due to contamination. Lymph node histology was suggestive of tuberculosis for 259 of the 268 patients. For six patients the histology was not suggestive of tuberculosis, and for three, no histology report was available. In these nine patients the lymph node culture yielded M. tuberculosis. Fifteen (17%) children and 26 (14%) adults had abnormal radiological features not suggestive of active pulmonary tuberculosis.
Response to treatment
Results at the end of treatment for 268 patients are shown in Figure 2. One hundred and sixteen (87%; CI 81–93%) of 134 patients in each treatment arm had a favourable response at the end of treatment. Four patients (3%; CI 0–6%) had an unfavourable response with the daily regimen; all were female, 18–22 years of age; in two, lymph node cultures for M. tuberculosis were initially positive, one of which was resistant to isoniazid. All four patients received between 90% and 100% of the prescribed treatment. One patient (1%; CI 0–2%) had an unfavourable response with the twice-weekly regimen; a male aged 13 years whose initial lymph node culture grew NTM. He received only 71% of the prescribed treatment. All five patients were re-treated for tuberculosis and responded well to treatment. Fourteen patients (11%; CI 6–16%) in the daily regimen arm and 17 patients (13%; CI 7–19%) in the twice-weekly regimen arm were classified as having a doubtful response.
Of the 21 patients with initial lymph node culture resistant to isoniazid, or streptomycin or both (Table 1), all eight treated with the twice-weekly regimen had a favourable response; of the 13 patients treated with the daily regimen, nine (69%; CI 44–94%) had a favourable response, three (23%; CI 12–34%) had a ‘doubtful’ response and one (8%; CI 0–16%) had an unfavourable response. All three patients with resistance to rifampicin (twice-weekly regimen), rifampicin and isoniazid (daily regimen) and rifampicin, isoniazid and streptomycin (twice-weekly regimen) had a favourable response.
Relapse of lymph node tuberculosis
In the 36 months following treatment, of the 232 patients who had a favourable response at the end of treatment (116 each treated with the daily and twice-weekly regimen, respectively), two died due to non-TB causes and one was lost to follow-up (Table 2). Three patients had relapses with lymph node tuberculosis: one treated with the daily regimen, at 21 months, and two treated with the twice-weekly regimen; at 10 and 20 months, respectively.
Table 2. Response to treatment and events during treatment and follow-up among 268 patients with lymph node tuberculosis admitted to a randomized clinical trial in Madurai, South India
Relapse/events during treatment and follow-up
Daily n (%)
Twice-weekly n (%)
*One each from ‘Favourable’ and ‘Doubtful’ groups.
† Two from ‘Favourable’ and one from ‘Doubtful’ group.
‡ Fisher's exact test.
§ Chi-square test with Yates correction.
¶ 126/134 = 94% and 129/134 = 96%.
At 36 months after completing treatment
Number assessed (‘Favourable’ plus ‘Doubtful’ at end of treatment)
Non-TB death/lost to follow-up
Relapse of lymph node TB
Lymph node events (new nodes/enlargement of nodes)
Adverse reactions to treatment regimens
Patients with any reaction
Of the 31 patients (14 plus 17 treated with the daily and twice-weekly regimen, respectively) for whom the clinical response was classified as doubtful at the end of treatment, 10 had repeat lymph node biopsies subsequently. Of these 10 patients, one in the daily regimen group had a lymph node histology suggestive of tuberculosis and a culture that yielded M. tuberculosis; this patient was treated as a relapse at 18 months. Six patients had a positive histology but negative cultures; one of these relapsed at 8 months (twice-weekly regimen). Three patients had negative histology and culture results. Apart from the two patients with relapses in this group of 31, the other 29 patients were doing well at 36 months after treatment.
Thus, in all five patients relapsed, two (2%; CI 1–3%) in the daily regimen arm and three (2%; CI 1–3%) in the twice-weekly regimen arm. Overall, 126 of 134 (94%; CI 90–98%) and 129 of 134 (96%; CI 94–98%) patients treated with the daily and twice-weekly regimen, respectively, had a favourable outcome at 36 months after completing treatment.
Fresh lymph node events during treatment and follow-up
Both during treatment and follow-up, new nodes appeared, or existing nodes increased in size in some patients in both regimens (Table 2). There were no statistically significant differences between the regimens. These nodes regressed in size spontaneously when treatment was continued and did not require change of treatment.
Adverse reactions attributable to anti-tuberculosis drugs
Adverse reactions possibly attributable to the anti-TB drugs were assessed in 276 patients admitted to the study excluding one for whom the initial diagnosis was incorrect (Table 2). Gastrointestinal symptoms such as nausea and vomiting were the commonest reported events. One patient treated with the twice-weekly regimen developed jaundice in the last month of treatment. Isoniazid and rifampicin were withheld and the jaundice subsided. All the other adverse reactions recorded were in adult patients and none in children. They ranged from mild to moderate and were managed with symptomatic measures.
Efficacy of test regimens
Both regimens tested in this study were successful both when assessed at the end of treatment and after 36 months of follow-up. The clinical response was favourable in 116 of the 134 (87%) patients in each treatment arm, at the end of treatment. Even though the response was classified as doubtful for 31 patients (i.e., they had palpable residual nodes that measured >10 mm in the maximum diameter), only two had a relapse of lymph node tuberculosis and the other 29 continued to remain well and free of symptom at 36 months after completing treatment. It has to be conceded that the measurement of the lymph node diameter was subject to assessment bias.
In all, only five patients (2%), two and three, respectively, in the two treatment arms had relapse of lymph node tuberculosis in the 36 months after treatment. Overall, 126 of 134 (94%) and 129 of 134 (96%) patients treated with the daily and twice-weekly regimen, respectively, had a successful outcome at 36 months after completing treatment. These results compare favourably with the experience of the BTS Research Committee, which found a 6-month daily regimen of rifampicin and isoniazid plus pyrazinamide for the first 2 months to be as effective as two 9-month regimens (British Thoracic Society Research Committee 1992; Campbell et al. 1993). Although the same drugs were used in our study, they were administered only twice a week, an important consideration in observed treatment. Intermittent regimens for tuberculosis are known to be as effective as daily regimens while being less toxic (Girling 1978). Further, intermittent therapy is cheaper and easier to observe. Thus, our regimen while using the same drugs as the BTS study, offers these additional advantages.
A study from Hong Kong showed no difference in efficacy among patients treated with 6 or 9-month fully intermittent regimens (Yuen et al. 1997). The intensive phase in this study was 4 months of four drugs that included streptomycin. Our results show that a less intensive regimen (three oral drugs in a 2 months, intensive phase) given only twice-weekly, produced comparable, if not superior results.
An encouraging feature of the 6-month daily regimen of rifampicin and isoniazid was that even patients with lymph node cultures resistant to one or more drugs fared well. Of 14 such patients, including one who was resistant to rifampicin and isoniazid, only one failed to respond to treatment while three had a doubtful response. Concerns have been expressed as to whether, in the scenario of increasing drug resistance, a two-drug regimen of isoniazid and rifampicin could be effective for the treatment of tuberculosis. While these concerns are justified, our study shows that at least for lymph node tuberculosis, where the bacillary load is small, satisfactory results may be obtained with the rifampicin–isoniazid regimen. Moreover, a similar regimen has been shown to be highly successful in treating patients for spinal tuberculosis (Indian Council of Medical Research/British Medical Research Council Working Party in South India 1989). However, in populations with significant isoniazid or multidrug resistance or where HIV seroprevalence is high this two-drug regimen may not be adequate. We did not test our patients for HIV infection. Preliminary results of a randomized clinical trial at the Tuberculosis Research Centre suggests that HIV positive pulmonary tuberculosis patients respond equally well to the standard 4-drug or 3-drug 6-month regimen compared to immunocompetent pulmonary tuberculosis patients, though the relapse rates may be higher.
Adverse drug reactions
Both the regimens were well tolerated. Adverse reactions attributable to anti-tuberculosis drugs were observed in only 17 of 276 (6%) of patients, almost all of them treated with the twice-weekly regimen. This is much lower than those reported among pulmonary tuberculosis patients treated with short-course chemotherapy, where usually four drugs are used in the intensive phase (Tuberculosis Research Centre 1983, 1986, 1997). Of patients treated with the daily regimen, only one had an adverse reaction, and probably reflects the fact that only two drugs were used in this regimen.
Physicians treating patients for lymph node tuberculosis often observe nodes enlarge in size or appear at new sites after treatment begins. This is often a source of concern, for both the patient and the treating physician. They have been widely reported and are thought to reflect immune reactions to the release of mycobacterial proteins caused by the bactericidal action of the treatment regimens (Campbell & Dyson 1977; British Thoracic Society Research Committee 1985; Anonymous 1987; Jawahar et al. 1990.). We also observed similar events among some of our patients and no statistically significant differences were seen between the two regimens.
Two striking observations from this study are the high ratio of females to males among the adult patients and the strongly positive tuberculin reactions. Although it is well known that lymph node tuberculosis, unlike pulmonary tuberculosis, is more common among females (Pamra & Mathur 1974; Farer et al. 1979; Research Committee of the Tuberculosis Association of India 1987), the female to male ratio of 5:1 that we observed in the 13–29-year age group was remarkable. Most patients had a very strong tuberculin reaction. This is unusual and we have not observed this type of strong reaction in other forms of extra-pulmonary tuberculosis (Rani & Rajeswari 2000).
Lymph node culture
Contrary to the common belief that mycobacterial culture from extra-pulmonary specimens is often not rewarding, lymph node specimens yielded positive cultures of M. tuberculosis in 178 of 265 (67%) patients, despite the fact that the specimens were transported from Madurai to Chennai, to the central laboratory 300 miles away. This involved a mean delay of 3–4 days between the time of the biopsy and the time of testing. Preservation of the specimens was facilitated by the use of multiple culture media and the use of a transport medium (Michison et al. 1983; Vanajakumar et al. 1997).
Strengths and limitations of the study
One of the strengths of this study is that we were able to follow-up all but three of the 268 patients who completed treatment, for 36 months to monitor for relapse. The limitation of our study, as in all clinical trials, is that the results may not be replicated under different conditions, unless adherence to treatment is meticulously ensured. This patient-management issue remains the biggest challenge in the fight against tuberculosis. A solution may lie in shortening the duration of treatment. The success of ofloxacin-containing 4- and 5-month regimens for treating patients with pulmonary tuberculosis, reported recently, is promising (Tuberculosis Research Centre 2002). A 3- or 4-month fluoroquinolone-containing regimen for the treatment of lymph node tuberculosis may be a realistic option in the near future.
Implications for TB control
The success of the two regimens reported on here has important implications for the tuberculosis control programmes. The WHO in its global efforts to control tuberculosis is promoting the widespread use of the directly observed treatment, short course (DOTS) strategy, in which lymph node tuberculosis is classified as category III, for which the recommended treatment is rifampicin and isoniazid for 6 months, plus pyrazinamide for the first 2 months, either daily or thrice-weekly (World Health Organization 1997). The intermittent regimen used in our study is the same regimen advocated by WHO, except that the drugs are given twice-weekly instead of thrice-weekly or daily. Our study shows that even this less-intensive (and less expensive) regimen can be highly effective.
However good the regimens are, unless they are acceptable for the patients they are unlikely to succeed. We feel that a twice-weekly regimen instead of a thrice-weekly one, and a two-drug regimen instead of a three-drug one are easier options that may appeal to patients.
In conclusion, the findings from our study show that lymph node tuberculosis can be successfully treated with a 6-month, twice-weekly regimen with three drugs in the intensive phase, or with a daily regimen of rifampicin and isoniazid. Both regimens were highly successful and well tolerated. With DOTS being expanded rapidly to cover almost 150 countries worldwide (WHO Report 2002), these regimens offer cost-effective and patient-friendly treatment options for countries planning to implement tuberculosis control programmes.
The successful conduct of this study owes much to the contributions of many people apart from the authors. We acknowledge the contributions of the following staff of the Tuberculosis Research Centre at Madurai: Dr. Palani Murugan and Dr. Paul Kumaran, Research Officers, Mr. Victor Mohan, Technical Officer, Mr. Subramanian K and Mr. Shripad Bhatt, Research Assistants (Statistics), Mr. Raja Sakthivel, Mr. Thiruvalluvan and Ms. Chandra S, Medical Social Workers, Ms. Arumaikannu Anbalagan, Public Health Nurse, Ms. Mary Eunice Chellammal and Ms. Chellam S, Clinic Nurses, Ms. Revathi V, Ms. Valarmathi R, Ms. Indirani V, Ms. Angel Nirmala, Ms. Pandeeswari, Ms. Rathinam M and Ms. Mohana, Health Visitors.
We acknowledge with gratitude the contributions of Dr. Renu Garg, WHO Consultant, Dr. Bandana Malhotra, and Dr. P R Narayanan, Director, Tuberculosis Research Centre, in the preparation of this manuscript in its various stages. The study was funded entirely by the Indian Council of Medical Research. No pharmaceutical company was associated with the study. Finally this trial would not have been possible but for the willing participation of the patients who constituted the subjects for this study. We acknowledge their contribution with humility and gratitude.