Risk factors for tuberculosis in dialysis patients: a prospective multi-center clinical trial

Profound alterations in immune responses associated with uraemia and exacerbated by dialysis, increase the risk for developing active tuberculosis (TB) after primary infection [1]. Frequent hospital contacts, older age and use of immunosuppressive drugs are additional factors explaining the higher prevalence of TB in these patients.

Tuberculosis in end-stage renal disease patients on regular haemodialysis (HDPs) presents a number of diagnostic challenges. In these patients the symptomatology of TB is often insidious and non specific, whereas the localization is often extrapulmonary [2, 3]. This makes the disease difficult to diagnose, delaying the initiation of curative treatment, which is a major determinant of the outcome [2‐4]. That is why annual skin testing with tuberculin skin test (TST) and chemoprophylaxis for all HDPs with a ≥10 mm response is recommended [5]. However, even today (with increased dialysis quality and dose) there is a high rate of anergy reported among HDPs and anergic HDPs are at increased risk of developing TB [6].

The aim of our study was to identify those HDPs who would benefit from TB chemoprophylaxis, using a combine approach of risk factors and common screening methods, like TST and chest X-ray. Specific-risks for TB were estimated as accordingly to TST results and chest X-ray findings, as well as history of TB, clinical important determinants, medication and specific dialysis related factors like dialysis efficiency and time on dialysis. Associating the later with the risk for TB we were also able to identify when the appropriate time to commence TB chemoprophylaxis in HDPs is.

Worldwide TB infection in dialyzed patients ranges from 5-25% and a 6.9-52.5-fold risk of TB is reported as compared to the general population [[3, 10, 11], and [12]]. These findings were verified in our study too. The absence of significant differences in Mantoux positivity and/or the incidence of active TB among the three medical centres involved in our study and the high incidence of extra pulmonary forms minimized the impact of clustering and/or the possibility of nosocomial spread.

Advanced age, diabetes mellitus and low BMI, are known factors of immunological imbalance, associated with TB development in the general population. In HDPs they have been also associated with depressed cell-mediated immune responses and anergy [6, 13‐17]. In our study, they were proved to be factors predisposing to active TB development in haemodialysis patients too. That is why they have to be considered evaluating for TB in the dialysis settings. Their existence supports the decision to prevent or even to treat when definitive diagnosis of clinical manifested TB is not possible, although there is strong presumptive evidence.

Most authors agree that to provide substantial protection against TB in HDPs, chemoprophylaxis should be commenced in all patients with a positive TST [5]. In the general population, TST remains the most satisfactory tool for diagnosing TB infection, and initial tuberculin sensitivity was well associated with the incidence of TB [18]. Boosted TSTs are usually recommended for immunocompromised patients like HDPs [19, 27]. Most authors agree that in the hemodialysis population two consecutive tests (within to a 7-14 days interval) with a higher strength (usually double) purified protein derivative, are sufficient to rule out the booster effect, maximizing the identification of the infected patients [20‐22]. However, in the current study, HDPs presented a weaker association between initial tuberculin sensitivity and the subsequent risk for TB. A substantial proportion of the HDPs that developed TB, had negative TSTs upon entering the study (14/24) and/or when the active disease was developed (8/24). As anergy distorted tuberculin reactivity, TST underestimated the incidence of TB infection among our patients [6, 23].

Unlike the TST, measurement of PPD reactivity by in vitro quantification of PPD-specific T cells has been reported unaffected by uremia-associated immunosuppression [24]. Nowadays, using a rapid 6-hour assay we can enumerate peripheral blood-derived interferon-γ secreting T cells responding to epitopes from an early secretory antigenic target-6 (ESAT-6) that is highly specific for M. Tuberculosis complex but absent from M. Bovis bacillus Calmete-Guerin (BCG) vaccine strains. Such whole-blood assays may thus be valuable alternative to skin testing in no endemic regions, but they are not still available for the National Health System of Greece [25, 26].

Obviously in HDPs, a negative TST has a comparatively low prognostic and diagnostic importance. However, as HDPs with positive TSTs presented increased risk for developing TB, a positive TST should never be underestimated.

Persons with abnormalities on chest radiographs consistent with prior, healed TB have a high risk for progression to active TB (2.0-13.6 per 1000 person-years of observation). Nodules and fibrotic changes may contain slowly multiplying tubercle bacilli with substantial potential for future progression to active TB [27]. HPDs with radiographic evidences of healed TB, demonstrated an almost 4-fold higher risk of developing active TB, independently of the result of the TST, even after allowing for the effect of the remaining risk factors. In our study the radiographic lesions were categorized and the specific significance of their appearance was assessed. Patients with non-calcified nodules and fibrotic changes in the upper lobes of the lungs, presented with the highest risk of TB. Bronchiectasis and pleural scarring are nonspecific findings that sometimes occur from previous pulmonary TB, but are more commonly caused by trauma or other infections. Conversely, calcified nodular lesions (calcified granulomas) and apical or basal pleural thickening posed a lower risk for future progression to active TB. In the case of HDPs such lesions may also represent metastatic calcifications attributed to the secondary hyperparathyroidism, a common clinical situation in uremic patients [28].

The weak correlation between chest radiography and TST positivity recorded in our, as well as in previous studies, verifies the value of chest X-ray in detecting TB infection in haemodialysis patients [29]. Commencing chemoprophylaxis in all HDPs with evidences of healed TB on chest X-rays, especially when diabetic nephropathy is an issue and/or when HDPs are malnourished and they have not been treated adequately in the past, regardless of their response to TST, will improve the control and prevention of TB in this high risk population.

An additional issue, investigated in our study, was the appropriate time for screening with TST and eventually commencing TB chemoprophylaxis in HDPs. There was a significantly higher prevalence and risk for TB recorded, in those patients dialysed for a shorter than a 12 month period. A result well associated with the depressed cell-mediated immune responses recorded in HDPs during an early dialysis stage [6]. It is important to notice that 16 out of the 24 TB cases developed before the first 12 months elapsed since undergoing dialysis, a finding verified in previous studies too [2, 30].

In our study, several unresolved issues concerning the identifications of those HDPs who would benefit with TB chemoprophylaxis, were investigated. Calcitriol, which is supposed to restore partly the lymphocyte function and low zinc levels that had been associated with higher anergy [13, 14] were not correlated with the development of TB. The efficiency of dialysis didn't influence the risk for TB either.

The above observations allow us to propose that in order to provide substantial protection against TB in HDPs, the annual skin testing which is recommended, has to be changed to testing upon admission to a regular dialysis program. HDPs with a positive TST and/or lesions suggestive of old healed TB in a recent chest X-ray are candidates for preventive therapy. As diabetic nephropathy and/or low BMI were proved to be independed factors of TB development, their existence supports also the decision to treat. Exception should be made for the elderly HDPs as they are more prone to the pharmacological toxicity of isoniazid [31]. However, as advanced age (> 65 years old) is an independed factor predisposing to TB, the need for a high index of suspicion and timely diagnostic work up should be reinforced in this group of HDPs. Tuberculin skin test and chest radiography upon entering a chronic dialysis program will provide a baseline with which to compare subsequent tests.