A controlled trial of short-course regimens of chemotherapy in patients receiving ambulatory treatment or undergoing radical surgery for tuberculosis of the spine.

Indian Council of Medical Research; British Medical Research Council.

Indian Journal of Tuberculosis; 1989; 36; 1-21.

Three hundred and four patients in Madras, South India, with a diagnosis of tuberculosis involving the thoracic or lumber spine were allocated at random to one of three treatment series. In two of the series the patients were ambulant from the start of chemotherapy and received isoniazid plus rifampicin for either 6 months (Amb 6) or 9 months (Amb 9). In the third series all patients had isoniazid plus rifampicin for 6 months and in addition underwent a modified `Hong Kong' operation of radical resection of the lesion and insertion of autologous bone grafts (Rad 6).

           After exclusions, the main analyses of this report concern observations for 3 years on 260 patients (85 Rad6, 83 Amb6, 92 Amb9). The clinical and radiographic characteristics of the three series on admission were broadly similar.

           At 18 months 58% of the 85 Rad, 58% of the Amb6 and 60% of the 92 Amb9 patients had a favourable status, i.e., they had no clinical evidence of active disease, in particular no sinus or clinically evident abscess and no myelopathy with functional impairment (defined as damage to or disease of the spinal cord secondary to the spinal lesion). In addition, they were capable of full physical activity, the lesion was quiescent radiographically and their allocated regimen had not been modified. At 3 years 68 (80%) of the Rad6, 72 (87%) of the Amb6, and 88 (96%) of the Amb9 patients had a favourable status. A further 8 (9%), 6 (7%) and 3 (3%) were classified as still not favourable solely because their disease was not yet considered quiescent radiographically. Of the 15 patients (9 Rad6, 5 Amb6, 1 Amb9) classified as unfavourable, 8 had a favourable status in all respects except that their allocated regimen had been modified. If such modifications are disregarded the proportion with a favourable status increases to 82% of the Rad6, 93% of the Amb6, and 97% of the Amb9 patients.

          There were 4 patients (all Rad6) who died from reasons associated with their spinal lesion.

          Major modifications to the allocated regimen were made in 11 patients (5 Rad6, 5 Amb6, 1 Amb9).

          Sinuses and/or clinically evident abscesses were present on admission or developed subsequently in 83 patients (22 Rad6, 33 Amb6, 28 Amb9). In the 51 patients (17 Rad6, 21 Amb6, 13 Amb9) who had a lesion on admission, resolution occurred more rapidly in the Rad series (p<0.001 at 2 months) but by 9 months the lesion had resolved in all except 1 (Amb9) patient. Sinuses and/or clinically evident abscesses developed after the start of chemotherapy in a higher proportion of the patients in the Amb series (27 of 175 patients) than in the Rad6 series (5 of 85 patients, p=0.05), these lesions resolving within 6 months in 20 and 4 respectively.

          Myelopathy was present on admission or developed subsequently (in all except 2 within the first 2 months) in 21 patients (6 Rad6, 6 Amb6, 9 Amb9), 16 (4 Rad6, 6 Amb6, 6 Amb9) of whom had functional impairment at some time. The functional impairment resolved on the allocated regimen in 11 patients (3 Rad6, 3 Amb6, 5 Amb9) and after operation in a further 4 (3 Amb6, 1 Amb9). In the 16th patient (Rad6) myelopathy with complete paralysis, which developed after operation, was still present at death.

           The rate of occurrence of complete bony fusion of the spinal lesion was similar in the 3 series, about a third of the patients having complete fusion by 12 months, and 60% by 3 years.

           The average total vertebral body loss on admission was 0.77 of a body in the Rad6, 0.85 in the Amb6 and 0.63 in the Amb9 series. By 3 years there had been a further loss of about quarter of a vertebrae in all 3 series, most of the loss occurring in the first 18 months. However, there was a difference between the series in the proportion of patients who showed an improvement (defined as a decrease in the total vertebral body loss of 0.25 of a vertebra or more), namely 5 of 79 patients in the Rad series but only 1 of 174 patients in the 2 Amb series combined (p=0.01).

            The most important conclusion from this study is that ambulatory chemotherapy with short-course regimens based on isoniazid and rifampicin is highly successful in the treatment of patients with spinal tuberculosis and, in terms of favourable status at 3 years, the most important assessment, the results for the ambulatory series were better than those for the radical series.


Back to List of publications / Home