Prevalence of latent tuberculosis in patients with hematological neoplasms in a cancer referral hospital in Mexico City

Tuberculosis has a major impact on the burden of disease globally. Approximately a quarter of the world’s population is infected with Mycobacterium tuberculosis [1, 2] and its distribution is mainly in low- and middle-income countries, such as Mexico [3]. In Mexico, in 2019, an incidence of 23 cases per 100,000 inhabitants was documented [4].

After exposure to the tuberculosis bacillus, the majority of immunocompetent people will achieve immunological containment, remaining in a silent state called latent tuberculosis (LTBI), characterized by the absence of clinical manifestations [1, 3]. However, under conditions of immunosuppression and/or in case of not receiving prophylactic treatment, it can reactivate and progress to “active” tuberculosis [1, 5‐7].

The vast majority of active tuberculosis cases are due to the reactivation of a primary infection [1, 3, 8]. In the general population, a patient with LTBI has a 10% risk of reactivation of tuberculosis, and half of these reactivations occur in the first 2 years of infection [5, 9]. In immunosuppressed patients this risk increases [9, 10], mainly in those with hematological neoplasms. For this reason, it has been recommended to screen for LTBI in immunosuppressed patients and to assess treatment in those who test positive, to reduce the probability of reactivation of tuberculosis [1, 5, 11, 12].

The most widely used diagnostic tool for the detection of LTBI is the tuberculin skin test (TST) [2, 3, 13]. Another method approved by international guidelines is IFNγ release assays (IGRAs), which offer greater specificity. However, both tests have decreased sensitivity in immunosuppressed patients. Detection algorithms have been proposed using the two tests consecutively to increase sensitivity or applying a TST booster. Until now there is little information reported in this regard in patients with hematological neoplasms. A recent Mexican study reported a 20% prevalence in bone marrow transplant candidates who underwent TST booster screening [14].

Treatment for LTBI has been shown to be effective in preventing its progression to active TB in immunocompromised patients [15‐18]. The currently recommended regimens are isoniazid (INH) 5 mg/kg once a day for 9 months [1], rifampin once a day for 4 months, or combined treatments (isoniazid plus rifapentine or isoniazid plus rifampicin for 3 months) [1, 9].

In Mexico, there are few publications on the prevalence of LTBI in patients with hematologic malignancies. The Instituto Nacional de Cancerología (INCan) has a large population of cancer patients. The aim of this study was to establish the prevalence of LTBI in patients with a recent diagnosis of hematological neoplasm and tolerance to treatment with isoniazid in our Institution.

INCan is a cancer, referral and teaching hospital located in Mexico City, which serves adult cancer patients who do not have social security, mainly from the central region of the country. It has 133 beds and a 40-bed hematology hospitalization service.

During the study period, 1372 patients were seen at the hematology department for workup, with a suspicion of hematologic neoplasms. In 96 of these patients (7%), a hematologic neoplasm was excluded. The TST was performed in 504 patients, which corresponds to 37% of all patients seen at the hematology department during the study period. Fifty-eight patients were excluded, in whom the diagnosis of hematological cancer was not confirmed. The final cohort for analysis were 446 patients (Fig. 1).

The diagnoses of the excluded patients were: non-hematological neoplasms (n = 31), lymphoid hyperplasia (n = 7), infectious pathology (n = 3) among which a patient with tuberculous splenitis stands out, a neoplasm was ruled out with no definitive diagnosis (n = 16); in one patient there were no available medical records.

Fifty-one percent of the participants were women. The median age was 54 years (40–66), 87.4% of the patients came from the central region of the country. Almost 50 % had no occupation (45.3%); 10.3% were illiterate, 13.5% did not complete elementary school, 59.9% received basic education and 17.2% had a higher degree.

Of the population studied, 40.1% had a history of smoking. The prevalence of diabetes mellitus was 15.9%, while 3.8% of patients were living with HIV. The hematologic neoplasms were: Non-Hodgkin’s lymphoma, 59.9% (n = 267); Hodgkin lymphoma, 17.9% (n = 80); multiple myeloma, 10.3% (n = 46); leukemia, 7.6% (n = 34) and other hematologic neoplasms, 4.3% (n = 19).

TST was positive in 139 of the 446 patients. The prevalence of latent TB was 31.2%. No active tuberculosis was documented at the time of hematological neoplasm diagnosis in any case. Table 1 shows the clinical and sociodemographic characteristics of the population studied by TST result.

Within the exposure history, 72% (n = 36) of the patients who reported a history of consumption of unpasteurized dairy products had a positive TST (OR 4.863, CI 95% 2.38–9.24, p < 0.0001). As expected, 74% (n = 17) of patients with a known exposure to TB patients had a positive TST (OR 5.074, CI 95% 1.928–13.349, p < 0.001).

On the other hand, only 6% (n = 1) of the 17 HIV positive patients of this cohort had a positive TST (OR 0.132, IC 95% 0.017–1.004, p = 0.029). Regarding the characteristics of the HIV patients, their median CD4 count was 164 cel/mm³; lymphoma diagnosis was an AIDS defining event for 41% of them, who were recently diagnosed and not on antiretroviral therapy at the moment of the TST.

Ninety-three patients received treatment for latent tuberculosis. Of these, 92 had a positive TST and it was decided to treat one patient with negative TST due to a history of direct exposure to a case of Tuberculosis. All patients with LTBI were treated with 300 mg of isoniazid daily, in addition to vitamin B6 for 9 months. Fourteen patients (15.1%) had an adverse reaction to INH: 7 had neuropathy, 4 had an increase in transaminases, and 3 had GI intolerance. A history of alcohol consumption was reported in three cases of neuropathy, one case of gastric intolerance and one case of transaminase increase. Two patients with neuropathy were smokers as well as one patient with gastric intolerance and one patient with transaminase increase. There were no associations between the exposure factors and the presence of adverse reactions.

Sixty-seven patients completed the treatment (72%). In 10 patients, there was no information available on whether or not INH was completed. Sixteen patients (17.2%) suspended treatment: 9 due to voluntary abandonment, 4 due to toxicity and medical indication (hypertransaminasemia (n = 3) and gastric intolerance (n = 1), one due to shortage of isoniazid. Active tuberculosis was documented in two patients; therefore, the regimen was changed to an antituberculous treatment with isoniazid/rifampicin/ethambutol/pyrazinamide.

The mean follow up time of the cohort was 1.77 years; 72% (n = 322) completed at least 1 year of follow up and 58% (n = 259) had at least 2 years (58%). The main reasons for not having follow up visits at one and 2 years were death (41 and 34% respectively) and lost to follow up (51 and 47% respectively).

This study reports a prevalence of latent tuberculosis in patients with hematological neoplasms of 31.2%. In Mexico, the prevalence reports vary widely according to the geographic area and population evaluated. A study of migrants at the border with the United States reported a prevalence of up to 55% using a combined strategy of TST + IGRA [19]. Another study carried out at a social security institution in Mexico City with diabetic patients, found a prevalence of 51% using only TST with an induration greater than 5 mm [20]. Another study reported a prevalence of 33.5% in patients with rheumatoid arthritis using TST booster [21]. C. Bourlon et al., found a prevalence of 20% in a population of hematology transplant recipients using TST booster. In our case, as the patients were evaluated before starting chemotherapy, the booster was not used. In their study, C. Bourlon et al. report a prevalence in healthy bone marrow donors of 40% which, together with the social security institution results, probably better reflects the prevalence in the general population of Mexico [14].

The TST has been the most widely used screening method for the diagnosis of LTBI. However, given its low specificity and lower sensitivity, in patients with a high risk of reactivation of tuberculosis, such as immunosuppressed patients, a combined IGRA and TST screening has been suggested. The above, in order to increase the diagnostic efficacy and in case of positivity of any of the tests, provide treatment to these patients [10, 12, 17]. In our study, within the group of patients with negative TST, no cases of active tuberculosis were documented. Although it was not designed to evaluate the diagnostic performance of the TST, as there were no cases of reactivation in the negative TST group, it could be considered that no further diagnostic tests are necessary to improve the sensitivity of the diagnosis.

Regarding immunosuppressed patients and those undergoing cytotoxic chemotherapy, it is important to reduce the probability of drug interactions and/or suspension due to intolerance to treatment. Currently, there are shortened regimens that are recommended as first-line treatment, such as INH/rifampicin dual therapy for 3 months [1, 9] for which a similar tolerance and efficacy was reported compared to the standard regimen [1]. It has been observed that with these regimens, there is greater adherence to treatment and a similar tolerance compared to monotherapy regimens [22‐25]. However, these regimens have not been studied in chemotherapy patients with the risk of interactions associated with rifampicin. In our study, 9 out of 93 patients (9.6%) abandoned treatment and this might probably be decreased with shorter schemes. More information is needed in patients with hematological malignancies in order to establish the safety and efficacy of shortened regimens and the risk of interactions with the use of rifampicin.

Most of the patients had adequate tolerance and adequate therapeutic adherence. Adverse effects occurred in 15.1%, but only 4.3% required the suspension of treatment. Neither one was a significant drug interaction demonstrated with the simultaneous use of different chemotherapy regimens during the follow-up period. Therefore, it could be considered, for now, as an adequate scheme for the treatment of LTBI in hematology-oncology patients in our institution and other similar scenarios.

The incidence of TB reactivation in our cohort was 0.45% in our cohort, with a population rate of 448 per 100,000, a very high rate and much higher than Mexico’s reported rate of 23 per 100,000 inhabitants in 2019 [4]. These results highlight the relevance of continuing screening and treating LTBI in high-risk populations like patients with hematological malignancies.

Another interesting fact that showed up in our study is INH and TST shortage, that can threaten TB prevention programmes as ours. This has been reported before in many countries and is a major obstacle for successful prophylactic programmes [26‐28]. Our data underscore the importance of screening hematologic patients in the context of a high prevalence of LTBI and high incidence of active TB in this population, and the difficulty to achieve it.

We also found an interesting association between LTBI and consumption of unpasteurized dairy products. This finding, related to infections caused by Mycobacterium bovis, is frequent in developing countries [29] and has been well documented in Mexico [30], where zoonotic tuberculosis caused by M.bovis has shown increasing trends and reported in up to 26% of isolates [31].

This study has some limitations. It is an observational and retrospective study, with only TST as the diagnostic method to assess LTBI; no other comparisons were available because IGRAS is not available in the INCan, as in many other public institutions in Mexico. Even when patients were followed during the period of greatest risk of reactivation, it is not possible to determine whether a longer follow-up could detect more cases of tuberculosis reactivation. Another limitation is that only 37% of all hematologic patients were tested. This low TST coverage is explained by the fact that implementation of LTBI screening in the Hematology Department started slowly in 2017, and at the beginning not all patients were evaluated for TST. Secondly, during the study period, there was a shortage of TST, and no screening was done during several months. Consequently, patients in this cohort were not chosen randomly but according to the availability of the TST, and we cannot guarantee that our results reflect the whole population of our institution. We do not have the description of all the characteristics of the haematologic patients that were not tested, to compare our cohort with, but we did compare the type of diagnosis distribution of the study group with the entire hematology population for 2017 and 2018 and patients distribution is very similar.

Despite these limitations, the close and standardized follow-up of the patients, as well as the cohort studied, provides important information about LTBI in patients with hematological neoplasms in a limited resources institution.

In conclusion, the prevalence of LTBI in patients with hematologic malignancies was 31.2%, which is within the range of prevalence reported in other similar populations in Mexico but lower than other non-immunosuppressed Mexican cohorts. This is likely due to a lower sensitivity of TST in immunosuppressed patients. However, no tuberculosis reactivations were observed in the negative TST group, suggesting that TST is a good screening tool. Isoniazid as monotherapy for 9 months was generally well tolerated with a relatively low proportion of significant adverse effects. It is important to evaluate the feasibility of shortened treatments in this type of patients to improve adherence to treatment.