INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bone marrow transplantation (BMT) is a unique model of extreme iatrogenic immunosuppression during which the recipient is susceptible to life-threatening opportunistic infections caused by a variety of pathogens. Several large series reviewing infections in BMT recipients have reported incidences of M. tuberculosis of < 0.1 to 2.2% (1), and there have been some isolated case reports (6). Because of the low rate of occurrence, no systematic analysis of tuberculosis in BMT recipients have been described. In Hong Kong, we have noted in our early BMT experience a higher incidence of tuberculosis than other countries (9). This provided us with an opportunity to study tuberculosis in our BMT recipients.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

One hundred eighty-three patients (17 yr of age or older) who underwent BMT at Queen Mary Hospital, Hong Kong, during a 4-yr period (1991 to 1994) were included in this study, and prospective evaluation was done up to the time of death or for at least 2 yr post-BMT.

The isolation and antimicrobial prophylaxis of BMT recipients have been described previously (10). Standard isolation facilities was applied, and prophylactic regimes for bacteria, fungus, virus, and Pneumocystis carinii were given using oral ciprofloxacin, amphotericin lozenge, acyclovir, and cotrimoxazole. Total gut decontamination was additionally used in partially matched sibling or matched unrelated donor (MUD) procedures.

Two groups of patients with evidence of old tuberculosis were considered for antituberculosis prophylaxis: (1) patients with old tuberculosis changes on chest radiograph (CXR) such as apical fibrosis or calcified granuloma, negative acid-fast bacilli (AFB) smear and culture, and no previous anti-tuberculosis treatment were given isoniazid (300 mg once a day) and rifampicin (450 mg once a day) from pre-BMT to at least 6 mo posttransplant; (2) patients with old tuberculosis changes on CXR, negative AFB smear and culture, and previous complete course of antituberculosis treatment were given ciprofloxacin (500 mg twice daily) for the first 3 mo post-BMT.

During follow-up, patients with clinical features suggestive of respiratory tract infection were assessed with chest radiographs and sputum examination. Endotracheal aspirates were collected in patients who were intubated. Bronchoalveolar lavage (BAL) was performed in selected patients who could not produce adequate sputum specimens. For the detection of acid-fast bacilli, respiratory secretions were processed for direct smear with auramine O stain and culture in Lowenstein-Jensen medium. All positive auramine O stain smears were confirmed by Ziehl-Neelsen stain. Cultures positive for acid-fast bacilli were identified by using the AccuProbe hybridization assay (Gen-Probe Inc., San Diego, CA). Antituberculosis drug sensitivity testing was performed by using the proportional method. All biopsy specimens were fixed and stained for acid-fast bacilli, apart from noting histopathologic features of tuberculosis.

A diagnosis of active pulmonary tuberculosis was based on: (1) the presence of M. tuberculosis documented in any of the respiratory tract specimen cultures, and (2) no positive cultures, but respiratory tract secretions or lung tissue histology showed acid-fast bacilli and other features of tuberculosis, and there was clinical and radiologic response to antituberculosis therapy. Remission of pulmonary tuberculosis was defined as resolution of CXR features and sputum smear and culture negative for M. tuberculosis for three times.

The stages of underlying diseases, duration of neutropenia, lung toxicity, acute and chronic graft-versus-host disease (GVHD) were defined and graded according to standard criteria (11, 12).

Data Analyses

Control patients were selected among the BMT recipients in the study period who did not develop pulmonary tuberculosis. For each patient with pulmonary tuberculosis, all patients without pulmonary tuberculosis who were of the same age and sex (n = 27) were included in the control group. A comparison of characteristics between these two groups of BMT recipients was performed to identify risk factors that were significantly associated with tuberculosis. The chi-square test was used for categorical variables, and Student's unpaired t test was used for continuous variables; p < 0.05 was considered as statistically significant. Multivariate analysis was not performed because of the small sample size (13).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Of the 183 BMT recipients under study, five required tuberculosis prophylaxis according to the criteria described. Three received isoniazid and rifampicin, and two received ciprofloxacin, respectively. In one patient a tuberculous pleural effusion was diagnosed 6 wk pre-BMT, and in another one 2 d pre-BMT in the conditioning period, and both were treated with four antituberculosis drugs through the transplant. None of these seven patients relapsed during follow-up.

Of the 183 BMT recipients, 10 (5.5%) developed tuberculosis after BMT. Their demographic and clinical characteristics are shown in Tables 1 and .2All 10 patients had received allogeneic (two were MUD) transplants. The median time of onset of symptoms attributed to tuberculosis was Day 150 (23 to 550), and the time between symptoms and definitive diagnosis of tuberculosis was mostly 6 to 7 wk, ranging from a few days to 7 mo in one patient. Most patients presented with fever and cough, and chest radiograph abnormalities were present initially in eight patients, with characteristic upper zone involvement in seven, and one patient had acute diffuse alveolar shadows and histopathology showed diffuse alveolar damage at autopsy. None of the patients had evidence of extrapulmonary disease. The respiratory tract specimens were smear positive for AFB in three patients, culture positive for M. tuberculosis in eight patients, and lung tissue histology was the diagnostic procedure in two patients. Five of the M. tuberculosis isolates were sensitive to the four first-line drugs isoniazid, rifampicin, streptomycin, and ethambutol, but one was resistant to isoniazid and streptomycin.

Patients in whom active tuberculosis was diagnosed were given the standard four-drug regimen in Hong Kong (isoniazid, rifampicin, pyrazinamide, and streptomycin) in the first 6 mo and two or three drugs for another 6 mo. In patients who could not tolerate oral medication, an intravenous regimen of rifampicin, ciprofloxacin, and amikacin was given until oral therapy could be instituted. Antituberculosis treatment or prophylaxis was continued until all immunosuppressive therapy ceased. The standard four-drug regimen was well tolerated by five patients, but ofloxacin was used as the fourth drug in two patients who could not tolerate the standard regimen because of hepatic impairment and low platelet count, respectively (14). No increase in side effects of antituberculosis drugs was observed. One patient did not receive any treatment because the diagnosis was made after death. Three patients died from other complications of BMT while having active tuberculosis, and six completed the course of treatment, although one subsequently died with no evidence of tuberculosis relapse.

A comparison of the BMT recipients with pulmonary tuberculosis and their matched control subjects is shown in Table 2. Allogeneic and MUD transplants and total body irradiation (TBI) were associated with the development of tuberculosis (p < 0.05, RR = 23.7 and 4.9, respectively). After excluding the three patients who developed tuberculosis or died before Day 100 and their corresponding control subjects, chronic GVHD was found to be associated with development of pulmonary tuberculosis (p < 0.05, RR = 3.6).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Little is known of pulmonary tuberculosis in BMT recipients. Excluding isolated case reports, there have been only five studies that reported a total of 10 cases among 5,193 BMT recipients (1). The overall incidence of 0.19% is far lower than the presently reported incidence of 5.5%. This finding supports the general observation that the immunosuppressed state would magnify infectious diseases that are endemic in the locality pulmonary tuberculosis in our population is at least 10 times greater than in other developed countries (15, 16). It is also in agreement with a previous report from our unit of high incidence of tuberculosis in a group of non-HIV, non-BMT immunocompromised subjects (17). Similarly, a BMT center in Spain, where tuberculosis is also endemic, reported an incidence of 2.2% (5), which is relatively higher than that reported from the United States and the United Kingdom.

Our analysis showed that chronic GVHD is highly associated with the development of tuberculosis. There are several possible mechanisms in chronic GVHD that may predispose the subject to tuberculosis. The normal host defense against tuberculosis is provided essentially by the macrophage-monocyte system with the T-lymphocytes as major coeffectors (18). In the BMT recipient, various elements of the human immune system reconstitutes itself postgrafting, modified by the influence of grafting complications and their treatment (19). There is early reestablishment of monocytic phagocyte functions followed in about 2 mo by repopulation of early lymphoid elements, and subsequently by emergence of T cells capable of regulatory functions a year later. The state of chronic GVHD itself confers a special susceptibility to infection via disruption of this reestablishment of normal host defense. Generally, patients who develop chronic GVHD have a marked delay of T-cell subset recovery. In particular, it has been shown that CD4⁺ cells, which are the major effector cells in cell-mediated immunity in tuberculosis (17), from recipients with chronic GVHD are low in numbers in proportion to CD8⁺ cells (20), and they fail to provide helper activity as compared with CD4⁺ cells from healthy recipients that constantly provide helper activity (21). The period of impairment of T-cell function may be indefinite in chronic GVHD. Because T cells play an important role in protective immunity against tuberculosis, chronic GVHD as a risk factor for development of tuberculosis post-BMT can be readily perceived.

It has been reported that PPD-specific memory T cells may be transferred from the marrow donor to the recipients of non-T-cell-depleted BMT and persist as part of the recipient's host defense (8). We have not performed the tuberculin skin test on the donors, although we expect that the majority, being adults among the indigenous population, have been exposed to infection either as a result of BCG or natural infection. Failure of acquiring adoptive immunity in those recipients with chronic GVHD may play a role in predisposition to TB. The treatment of chronic GVHD with immunosuppressive drugs is another factor that may be important. However, our analysis did not show any significant difference of steroid usage in the group who developed tuberculosis and the control group. It is possible that the treatment effect on GVHD outweighs its risks in this regard.

All of our patients who developed pulmonary tuberculosis underwent either allogeneic or MUD transplant. This is in keeping with the previously reported incidence of tuberculosis of approximately 0.43% in allogeneic BMT (1, 5, 8) and 0.08% in autologous BMT (3, 5, 8). The high association of tuberculosis and allogeneic/MUD transplant was likely in part due to the high incidence of chronic GVHD after these procedures.

Total body irradiation was also shown to be a risk factor associated with the development of pulmonary tuberculosis. Besides imposing additional immunosuppression on the host in general, TBI also hampered normal function of alveolar macrophages, which are essential to the defense against M. tuberculosis (18). Total body irradiation has also been shown to be an independent risk factor for the development of aspergillosis in BMT recipients (10), another infection that is highly dependent on alveolar macrophage function for defense.

The lack of a statistically significant relationship between the development of pulmonary tuberculosis and acute GVHD could be related to the shorter duration of acute GVHD. Patients without latent tuberculosis did not have the chance to be exposed to infectious cases since they were still hospitalized in isolation units, but subjects who have a latent focus may experience reactivation early in the post-BMT course. Both patients who developed tuberculosis early post-BMT had severe acute GVHD. Moreover, recipients with severe acute GVHD may have died from overwhelming GVHD or sepsis, thus masking any potential to develop tuberculosis.

The clinical and radiologic presentation suggested that the tuberculosis was reactivation in nature rather than primary infection. This was compatible with the high incidence of previous exposure relating to the uniform policy of BCG vaccination at birth and the high prevalence of tuberculosis in the general population in our locality. Logically, one would expect a higher incidence of extrapulmonary tuberculosis since the patients' cell-mediated immune responses were suboptimal in localizing the infection. The sites of organ involvement were specified in eight of the 10 reported cases in the literature, of which six were pulmonary, one was disseminated, and one was meningeal infection. Paradoxically, none of our 10 patients had extrapulmonary or disseminated tuberculosis. This was probably because of our practice of close scrutiny and the high index of suspicion with prompt institution of antituberculosis therapy pending subsequent confirmation in suspected cases.

The practice of using tuberculin skin testing to identify at-risk subjects for preventive therapy has not been utilized because of the difficulty in interpretation of a positive tuberculin skin test in an area of high infection prevalence, such as Hong Kong (22). Among BMT subjects, this would be further complicated by their varying degrees of anergy. The seven subjects receiving preventive therapy or active treatment at the time of BMT did not have any relapses, suggesting that preventive treatment was efficacious.

The overall response to treatment with the standard four-drug regimen for a longer duration (1 yr as compared with 6 mo in immunocompetent subjects) has been satisfactory. The absence of relapse after termination of treatment suggested that secondary prophylaxis would not be necessary as long as immune function has been restored.

In summary, our experience suggests that pulmonary tuberculosis is not a rare opportunistic infection in BMT recipients in our locality where the background prevalence of tuberculosis in the community is relatively high. We have also identified certain risk factors associated with development of tuberculosis, including allogeneic transplantation, use of TBI, and the presence of chronic GVHD. With the rising incidence of tuberculosis in countries that previously enjoyed a very low prevalence of the disease (16), attributed to the growing population of HIV-infected subjects with tuberculosis, and the changing patterns of population migration, it is important to bear a high index of suspicion of M. tuberculosis as a pathogen in marrow transplant recipients.