Treatment outcome of extra-pulmonary tuberculosis in Finland: a cohort study

The method of identifying all culture-confirmed tuberculosis cases in Finland, with the first positive culture sample date between January 1^st, 1995, to December 31^st, 1996 (N = 1059), present in either the National Infectious Disease Register (NIDR) or acquired through a separate query to all microbiological laboratories has been described elsewhere [21].

A case of non-pulmonary tuberculosis was defined as culture finding for M. tuberculosis in other specimen type than respiratory secretion, in the absence of M. tuberculosis in culture from respiratory secretions and sputum smear positivity for acid fast bacilli [18, 22]. With this definition, 322 (30% of the whole cohort) constituted the base-frame for the extra-pulmonary tuberculosis cohort.

Out of the 322 cases in the extra-pulmonary tuberculosis cohort, complete medical records were available for 311 (96.6%) (Figure 1). Among these 311 cases, four (1.3%) had previously been treated for tuberculosis after the year 1970, and were excluded from the outcome analysis as re-treatment cases. We further assessed the detailed clinical information in patient charts in order to divide the cases as deep and superficial disease. During this assessment we identified sample types not directly classifiable as pulmonary or non-pulmonary. All the patient charts and autopsy reports of this group of cases were assessed. We found 27 cases as having been erroneously classified by our original algorithm as extra-pulmonary TB. These cases were diagnosed from the culture of gastric aspirate or biopsy samples from lung tissue. In four cases the information was not sufficient for judging the location of the disease. These 31 cases were excluded from the extra-pulmonary study group leaving 276 extra-pulmonary cases with no previous treatment documented (Table 1). These extra-pulmonary TB cases were divided to deep and superficial disease based on the site of bacteriological sampling and/or ICD-codes. Lymph node and skin tuberculosis were considered as superficial and other forms as deep extra-pulmonary disease. If the same patient had both deep and superficial disease, the case was classified as deep disease.

Tuberculosis treatment was initiated in and supervised by the pulmonary departments of public hospitals in the great majority of cases. Chemotherapy combination was decided by the treating physician. In our retrospective record review, the combination actually given to each patient was grouped into six categories (Table 2) in order to describe the variability in treatment combinations used. Definitions for the grouping were based on the national recommendations in Finland following the recommendations by WHO, ATS and BTS [23‐28]. We have previously described the treatment grouping in detail [18]. Pauses of chemotherapy were recorded only when lasting at least one week, and calculated only for standard treatments.

The categories of WHO/EuroTB recommendation for outcome monitoring are cure, treatment completed, failure, death, treatment interrupted (default), transfer out, and on treatment at 12 months [5]. The duration of the follow up period is defined as 12 months from the beginning of the treatment or the date of diagnosis, and the first outcome registered as final.

The WHO/EuroTB category 'treatment interrupted' includes all interruptions, whether caused by a patient or by a treating physician. For analysing these two separately, we divided 'treatment interrupted' into 'physician's decision to stop early' and 'default' for interruptions due to patient only [18]. The outcome was recorded as 'death', if the case died before starting the treatment, during the treatment, or the date of death was within 14 days after cessation of the anti-TB drugs.

The outcome categories used are listed in Table 3. The outcome was categorized as favourable in cases of cure and treatment completed, and as unfavourable in cases of failure, death, default, physician's decision to stop early and transfer out.

A case was defined as immigrant if the country of birth was not Finland or, in the absence of country of birth, the nationality was other than Finnish. Immunosuppressive treatment was defined as corticosteroid treatment (>40 mg per day of any duration, or any daily dose with treatment duration exceeding one month), cytotoxic or cyclosporine treatment, or radiation therapy during the preceding year. For inclusion in the group of social risk factors, a case had a history of alcohol abuse, unemployment, imprisonment or homelessness recorded in patient records. To be included in the diabetes risk group the case had juvenile or adult onset disease on medication.

Specialty responsible for treatment was that of the unit treating patient as an inpatient or outpatient.

The ethics approval for this study was acquired from the National Research and Development Centre for Welfare and Health.

Median age comparisons where tested using Median test. For comparison of proportions, Chi-square or Fisher's generalised exact test with Monte Carlo simulation was used. Stata 11.0 (StataCorp LP Texas, USA) and IBM SPSS for Windows, version 18.0.2 (Chicago, IL, USA) was used for statistical calculations. We used binary and multinomial logistic regression model to assess the relationships between all predictors that were found to be significant in pulmonary tuberculosis (to compare extra-pulmonary tuberculosis to pulmonary tuberculosis in respect to these variables), and other variables which in the univariate analysis had a p-value less than 0.20, with binomial and trinomial outcome variables, in which the reference class was favourable outcome. To the final model, predictors were selected by forcing all those variables that were significant in pulmonary tuberculosis, and from other variables selection was by forward stepwise method. P-values under 5% were considered as significant. For univariate results also Chi squared test and Fisher's exact test were used.

Tuberculosis chemotherapy was given to 257 (93.1%) of the 276 cases. Half of the cases (140/257; 54.5%) received standard treatment A, 28 (10.9%) received standard treatment B, and 30 (11.7%) received standard treatment with additional drugs (D) (Table 4). In 18 (7.0%) of the cases the intensive phase was short (category C). A total of 41 cases (16.0%) received either other combinations of tuberculosis drugs or ineffective treatment. Most of the cases (177/257, 68.9%), were treated with the same combination during the whole treatment. Nineteen (6.9%) of the 276 cases were untreated, the majority (16; 84.2%) of whom died before the diagnosis of tuberculosis was established.

The median age differed significantly between the treatment groups; it was 70,5 (range 6,9-94,4) years for category A, 72.6 (18,5-90,4) for category B, 58.2 (6,6-91,0) for category C, 57.5 (15,1-88,2) for category D, 72.6 (60,8-71,6) for other combinations and 67.8 (18,6-94,4) for ineffective treatment (p = 0.005). There were 62 (44.3%) men in treatment group A, 10 (35.7%) in B, 6 (33.3%) in C, 15 (50.0%) in D, 8 (22.2%) in other combinations and 4 (80.0%) in ineffective treatment group. The difference was significant (p = 0.04). There was no statistical difference between groups regarding the presence of a social risk factor (p = 0.32).

A favourable outcome was achieved in 157/276 (56.9%) of the cases, consisting of those cured (8.0%) and treatment completed (48.9%). The proportion with favourable outcome varied between 35.7 to 80.0% in different treatment groups. There were no treatment failures in the cohort.

The proportion of cases defaulting or transferring out was 14/276 (5.1%), and of them 5/14 (35.7%) were immigrants and 2/14 (14.3%) had a social risk factor. In 31 (11.2%) cases treatment was stopped prematurely by physician. Death was the outcome in 17.4% (48/276) cases, including cases not treated.

The proportion of cases with favourable outcome differed significantly between treatment groups (P < 0.001), and was smallest in treatment groups B and D (35.7% and 43.3%). The proportion of deaths differed between groups, and was lowest in treatment group D (3.3%). In the ineffective treatment group the proportion of deaths was higher (20.0% versus 12.1%) than in the other combinations of TB drugs.

24 of the cases were still on treatment at 12 months, and in two cases it was not possible to judge the outcome from the information available. Of the remaining 250 cases, 19 were not treated, among them 11 (57.9%) were men and 8 (42.1%) women. Of this group, with a median age of 81.0 years, 16 died before treatment and 3 were left untreated. These 45 cases were excluded from the analysis of risk factors of poor treatment outcome.

In univariate analysis, patient-related risk factors which were significantly associated (p < 0.05) with death, were high age, immunosuppression and diabetes (Table 5). Treatment system -related risk factors which were significantly associated with death were the specialty of the treating unit (internal medicine, general medicine in primary care), change of specialty responsible for treatment and less than five treated cases per year per unit (Table 6). There was no statistically significant patient-related personal risk factor association for other unfavourable outcome (i.e. transfer out, default or physician's decision to stop early) (Table 5). Treatment system -related significant risk factors for other unfavourable outcome were standard treatment B, other treatment combination, change in treatment group and pause(s) during treatment (Table 6). For all unfavourable outcomes together, i.e. death and other unfavourable outcome combined, significant predictors in univariate analysis were high age, immunosuppression, specialty of the treating unit (general medicine in primary care), change of specialty responsible for treatment, less than five treated cases per year per unit, standard treatment B, other treatment combination and pause(s) during treatment.

Significant independent risk factors for death in multinomial logistic regression model were male sex, high age, immunosuppression, any other than a pulmonary specialty being responsible at the end of the treatment and other than standard combination of treatment. For other unfavourable treatment outcomes, significant risk factor was treatment group B (Table 7). Deep site of TB was inversely associated with the risk of other unfavourable outcome (p = 0.02).

For death and other unfavourable outcomes together, significant risk factors in the binary logistic regression model were high age (p = 0.02), immunosuppression (p = 0.001), treatment group B (p < 0.001) and any other than a pulmonary, internal and general medicine specialty being responsible at the end of the treatment (p = 0.01). Almost significant risk factors were male sex (p = 0.06) and specialty of the treating unit (general medicine in primary care) (p = 0.06). Deep site of tuberculosis was almost significantly protective (p = 0.07) (Table 7).

When we analysed all 250 cases in the cohort with known outcome, including those 19 without treatment, the associations observed in multivariate analysis as significant were the same as with the 231 cases of the presented analysis.