During the course of effective treatment for tuberculosis (TB), patients may experience unexpected deterioration. Though sometimes difficult to differentiate from clinical failure, these paradoxical reactions are generally defined as clinical or radiologic worsening of preexisting TB lesions, or the development of new lesions in a patient who initially improves, that are not attributable to the normal course of disease.1–3 It is important to recognize and differentiate paradoxical reactions from drug resistance or infection with other organisms, for which different management strategies are required.
The paradoxical response is sometimes termed as the immune reconstitution inflammatory syndrome, which reflects the immunopathologic basis of many of these episodes. Immune reconstitution inflammatory syndrome has been most frequently reported among human immunodeficiency virus (HIV)-positive adults receiving antiretroviral therapy4,5 and has been reported in more than one-third of patients coinfected with TB,5,6 with worsening radiographic changes after initiating antiretroviral therapy noted among nearly one-half of such patients.7 The clinical spectrum of paradoxical responses has not been well studied among HIV-negative children and adolescents; the only pediatric case series was published in 1955, followed by individual case reports.8–10
To examine the incidence, timing of, and risk factors for deterioration, we studied a cohort of HIV-negative children treated for TB at our center during a 7-year period. All patients had very close follow-up for at least 1 year after therapy. We also illustrate the clinical challenges in 1 patient. To our knowledge, this is the largest study since 1955 of deterioration and paradoxical reactions in HIV-negative pediatric patients treated for TB and the first to explore risk factors for deterioration.
A previously healthy 5-month-old boy was referred to the TB program after his father was diagnosed with fully drug-susceptible pulmonary TB. The child was asymptomatic, with a negative Mantoux test and normal chest radiograph. Isoniazid prophylaxis (12 mg/kg/d) was commenced. Within the week, he developed a cough. Repeat imaging showed a right middle lobe infiltrate. After admission and 3 gastric aspirates, rifampin and pyrazinamide were added to the regimen. Serologic results for HIV were negative, 1 of the gastric aspirates was smear positive for mycobacteria and grew Mycobacterium tuberculosis. After almost 3 weeks of directly observed therapy (DOT), he was readmitted with fever, acute-onset stridor, and decreased breath sounds over the right hemithorax. Computed tomography of the chest demonstrated large multilobulated hilar and paratracheal lymph nodes causing significant narrowing of the distal trachea and bronchus intermedius. Prednisone therapy at 2 mg/kg/d for 14 days was started, with symptomatic improvement within 48 hours. No additional agents were added, and the child was well with a normal chest radiograph result after 6 additional months of therapy and in follow-up for 3 years.
Study Population and Setting
We identified all patients with active TB who were followed at The Hospital for Sick Children Tuberculosis Clinic in Toronto, Canada, between January 2002 and July 2009, by review of discharge diagnoses and the TB service registry. The clinic manages 15 to 20 new cases of active TB disease per year and is the primary site for the evaluation of pediatric TB in Toronto. All patients with tuberculous disease are tested for HIV at the time of diagnosis. During the study period, patients with TB disease were managed using standardized protocols.11,12 Patient data were recorded in a standardized database form at initial visit and included a TB skin test, blood work for complete blood count, and renal function and liver enzymes and function tests. Chest radiography was also performed and the radiograph read before the end of the clinic visit, and other diagnostic investigations were performed as per the chief complaint. Patients were seen at least monthly by a consistent clinical team with 1 supervising clinician. Chest radiographies were performed at the end of therapy in the patients with pulmonary disease and where clinically indicated. The clinic protocols use daily antituberculous therapy. DOT is performed by Toronto Public Health staff on 5 of the 7 days of the week. The diagnosis of TB disease was established in accordance with the World Health Organization's Tuberculosis Standards case definition.13 Where there was no positive culture for M. tuberculosis, intrathoracic TB was diagnosed by chest radiographic abnormalities consistent with TB and either a history of exposure to an infectious case or origin in or a visit to a high-burden country and evidence of TB infection by tuberculin skin test. Patients with suspected extrapulmonary TB had specimens obtained from suspected sites of involvement—where there was no culture or growth, the diagnosis was based on epidemiologic features consistent with TB exposure, consistent signs and symptoms, findings on diagnostic imaging, and a positive tuberculin skin test. Wherever possible, our program obtained cultures from children and adolescents with suspected TB disease.11 However, in children with radiographic abnormalities such as hilar lymphadenopathy and known close contact with an infectious source case, initial cultures were not always obtained and source case susceptibilities were then used to guide therapy. We extracted information from patient records, which include data routinely recorded at the first TB clinic visit.
Unexpected deterioration was defined as clinical and/or radiographic worsening of preexisting TB lesions or the development of new lesions associated with clinical and/or radiographic changes 10 days or more after the start of antituberculous therapy after demonstrating initial clinical response.1,3,14 Patients were included if they presented with worsening or new disease on assessment of symptoms in the outpatient clinic or emergency department. In the absence of recorded new symptoms, we also included those with documented radiographic improvement followed by new radiographic changes detected beyond the fourth month of therapy. This long interval was chosen because chest radiographs in childhood TB may temporarily worsen during the course of therapy.15 The diagnosis of unexplained deterioration required the absence of evidence of disease relapse caused by poor adherence to TB treatment or drug resistance, or another diagnosis, including bacterial or viral infection.2,16 Drug-susceptibility testing was performed on all cultures positive for M. tuberculosis at initial diagnosis at the regional public health laboratory. For patients who deteriorated, we ascertained sites of disease at diagnosis and at deterioration, and reviewed laboratory, microbiologic, radiographic and histopathologic reports. Where a patient had multiple sites of deterioration, each site was counted as a separate episode of deterioration.1 Actions at the time of deterioration were recorded, including observation, initiation of a course of steroids, and/or modification of antituberculous therapy.
Data were analyzed using SAS statistical software (SAS Institute, Inc., Cary, NC), version 9.1. Medians were calculated using Excel 2000 software (Microsoft Corp, Redmond, WA). The outcome of interest in this study was deterioration on antituberculous therapy. Potential risk factors associated with deterioration were analyzed by univariate logistic regression: age, weight, sex, and location of primary disease. Categorization of number of sites involved was as 1 or >1; age as 0 to 2 years, 2 to 12 years, and >12 years of age; and weight as ≤25th percentile for age or >25th percentile for age. Categoric data were examined by the χ2 test or Fisher exact test for smaller sample sizes. Statistical significance was defined as P < 0.05. In consideration of the low deterioration proportion and the small overall sample size, variables with a P < 0.25 on univariate analysis were examined by exact multiple logistic regression. The study was approved by the Research Ethics Board at The Hospital for Sick Children.
A total of 112 children and adolescents were identified with clinical and/or microbiologic diagnosis of TB during the 7-year period. Two patients with HIV were excluded from the study; neither of them deteriorated during the course of therapy.
The median age of the remaining 110 patients included in the study was 11.8 years (range, 2 months–17 years). Of them, 61 (55%) had cultures positive for M. tuberculosis. The diagnostic criteria for the remaining 49 patients are shown in Table 1. Of the culture-negative patients, chest radiograph abnormalities, chiefly hilar or mediastinal lymphadenopathy, were found in 39 (80%) during evaluation of contact with an infectious source case.
Of the 110 patients, 62 (56%) were foreign born. Of the 48 Canadian-born children, 46 had history of travel to an endemic area or known contact with an infectious source case. Exposures, defined as birth in, travel to, or contact with a person from an endemic region, were associated with South East Asia (30%), the Western Pacific (24%), the Eastern Mediterranean (20%), Africa (12%), the Americas (10%), and Europe (3%). Intrathoracic disease (involving mediastinal lymph nodes, pleura, and/or lungs) was found in 75 (68%) patients; 16 (15%) exclusively had extrathoracic disease, and 19 (17%) had multiple sites of involvement. Thirty-eight patients (34%) were ≤25th percentile for weight, with one-half of this group below the fifth percentile. None was lost to follow-up in the duration of therapy.
A total of 19 patients (17%) experienced clinical or radiographic deterioration after initiating TB therapy. Of these, 4 were excluded from further analysis because of identifiable causes, namely stitch abscess, sickle cell-related vaso-occlusive crisis, multidrug-resistant strain of TB, and poor compliance with DOT visits. Among the 15 remaining patients (14%) who deteriorated (Table, Supplemental Digital Content 1, https://links.lww.com/INF/A993), none had an identified underlying immunodeficiency, and all had been receiving DOT. The median duration of therapy before deterioration was 80 days (range, 10–181 days). Eleven patients (73%) deteriorated more than a month into therapy. Six patients experienced deterioration of multiple sites, for which each site was accounted as an episode. In 22 episodes of deterioration, the most common site of involvement was intrathoracic lymphadenopathy with or without parenchymal involvement (7), followed by extrathoracic lymphadenopathy, including intra-abdominal, inguinal and cervical (4), abdominal viscera (2), brain (1), bone (1), pericardium (1) and pleura (2). Four patients developed symptoms at sites remote from the site of primary disease, all of whom were initially asymptomatic. Of these patients, 2 (cases 14 and 15, Table, Supplemental Digital Content 1, https://links.lww.com/INF/A993) had negative initial radiographic evaluation results of the sites of deterioration.
The management of the deteriorations was variable, depending on severity of clinical presentation. Treatment for suspected bacterial infection was commenced in 3 patients, 2 of whom had new-onset respiratory distress and 1 who presented with an acute abdomen. In none of these instances was a secondary bacterial infection ultimately identified. Nine patients had changes in antituberculous medications: in 4 patients, ethambutol and/or pyrazinamide were reinstituted in addition to their regimen of rifampin and isoniazid, and in 5 patients, second-line agents—a fluoroquinolone and/or amikacin—were added to the therapy. Of these 9 patients, 7 had appropriate specimens collected to exclude other diagnoses. Corticosteroid treatment was initiated in 9 patients and restarted or increased in further 2 patients who had received corticosteroids at the commencement of antituberculous therapy. Indications for corticosteroid therapy included increased intrathoracic lymphadenopathy with airway compression (4), new abdominal masses or strictures (2), worsening pleural and/or abdominal disease (2), and pericardial effusion (1). The median dose of prednisone was 1 mg/kg/d (range, 0.5–2 mg/kg/d), with an average duration of 14 days (range, 5–30 days) before tapering. One infant with miliary TB (case 9) required corticosteroid therapy for almost 1 year.
All patients responded with clinical and radiographic improvement and were clinically and radiographically cured, with follow-up until at least 2 years after therapy. One patient required ileocolic resection because of stricture formation with bowel obstruction. All but 4 patients returned to baseline. Sequelae included hemiparesis in a patient with TB of the central nervous system following cerebrovascular episodes (case 1), reactive airway disease in 2 patients with pulmonary and miliary TB (cases 5 and 9), and vertebral collapse in a patient with spinal TB who was not a candidate for corticosteroids because of to poorly controlled type 1 diabetes (case 12). No secondary drug resistance was identified.
To investigate possible predictors of deterioration, we compared those patients who developed paradoxical reactions with those who continued to improve during the course of antituberculous therapy (Table 2).
From univariate analysis, involvement of more than 1 site at diagnosis and weight-for-age below the 25th percentile were significantly associated with deterioration. Of the patients who deteriorated, 9 (60%) were at or below the 25th percentile for weight, as compared with 29 patients (31%) who remained well. A higher proportion of children less than 24 months of age deteriorated as compared with older populations, but this did not attain statistical significance.
One-third of children with extrapulmonary lymphadenopathy deteriorated during therapy, suggesting a 3-fold increased risk that trended toward statistical significance. Although our sample size did not permit detailed comparisons, there appeared to be no increased risk of deterioration with intrathoracic disease, namely parenchymal involvement or mediastinal lymphadenopathy. There was no statistically significant difference with respect to sexes, involvement of different anatomic sites, or country of origin between those who deteriorated and those who remained well during therapy.
To our knowledge, this study is the largest series in an era of optimal antituberculous therapy to identify and determine the incidence of unexplained deterioration in an immunocompetent pediatric population. We found that deteriorations are not infrequent, and, as illustrated by the management of these patients at the time of deterioration, these events may be clinically difficult to differentiate from disease progression, relapse, and drug fever or secondary complications. Characterization of these reactions is important for better clinical recognition and management.
Our study included children from inpatient and outpatient settings with a variety of clinical presentations. Strengths of the study included no loss of patients to follow-up during the study period, excellent links to public health staff that followed the patients closely between clinic visits with DOT, and review of all patients who presented with worsening clinical or radiographic disease, accurately reflecting events among these patients during therapy. It may be that the frequency of deterioration in our series was affected by referral bias, with more ill patients who might be at increased risk of deterioration being treated at our center rather than at other health care facilities in the Greater Toronto Area. However, our center is the main treatment site for pediatric TB disease in Toronto, and local health units try to refer the care of all children and adolescents with TB disease and close contacts of infectious adults to our clinic. We have also previously shown that the pattern of adolescent TB disease found in a population-based study in the province of Ontario17 was similar to that found in our center.18
The incidence of unexplained deterioration in our patient population was 14%. Paradoxical responses have been reported among HIV-negative adults being treated for TB,1,19,20 with an incidence between 2% and 30%. However, most of the literature comes from studies of HIV-positive adults who deteriorated during antituberculous therapy, with retrospective studies between 1998 and 2008 reporting an incidence of 8% to 43%,21 and 41% in a subsequent prospective cohort.22 There are no cohort studies of HIV-positive children who deteriorated during antituberculous therapy.
In our study, patients experienced deterioration between 10 days and 6 months into therapy. Although the earliest report suggested that the highest frequency was seen within 2 weeks,8 recent studies among HIV-negative adults have noted deteriorations at a median of 60 days (range, 20–109),16 87 days (range, 23–157),2 and 46 days (range, 10–405 days)3 after starting therapy. In our series, neither age nor extent of disease seemed to be associated with earlier reactions.
In our study, intrathoracic lymph nodes were involved in 7 of 22 episodes (31%) and extrapulmonary lymph nodes in 4 episodes (18%). Lymph node enlargement is a relatively common occurrence during therapy, with the incidence of 4% to 13% reported among HIV-negative patients and 14% among HIV-positive patients.2,23 Worsening clinical status due to mediastinal lymphadenopathy was described in all the deteriorations in the 1955 pediatric series.8 We were unable to assess other sites as predictors of deterioration, although central nervous system involvement has been reported in up to 49% of episodes as new or worsening tuberculous meningitis, intracerebral space-occupying lesion, or radiculomyelopathy.1,9,24,25
It may be difficult to distinguish unexpected deteriorations from drug resistance, treatment failure, or infection with other pathogens, and thus may lead to a change in therapy. A short course of steroids in the setting of unexplained deteriorations is often beneficial,2,5,26,27 and the response to the antitumor necrosis factor antibody, infliximab, in a steroid-refractory case28 further supports an immunopathogenic process. In our series, most patients responded to a 2-week course of prednisone at 1 mg/kg/d, with increased dose or duration of therapy for those with more severe disease at presentation. However, patients may also require surgical intervention by way of drainage of effusions or abscesses or resection of abdominal viscera. An escalation of antituberculous therapy could be considered if the susceptibility profile of original cultures were unavailable, given the risk of multidrug-resistant TB.
Univariate analyses to assess predictors of deterioration demonstrated that those patients who worsened during therapy tended to be below the 25th percentile for weight (P = 0.03). It is plausible that, during the course of antituberculous therapy, improvements in general or specific nutrient levels may lead to immune reconstitution and the development of paradoxical responses, as have been found among malnourished HIV-positive children29 and HIV-negative adults.19
Our analysis also suggests that patients with multiple sites of disease at presentation were 4 times more likely to deteriorate during therapy than those with localized disease. Earlier studies have documented an association between deterioration and disseminated or extensive disease,2,5,8,23,27,30 which may elicit a more severe inflammatory response following effective therapy because of its higher antigen load. The differences in age-related incidence of deteriorations did not attain statistical significance but is an important area for examination in a larger study, given the known differences in the immune responses to TB childhood.31,32
Our study is the largest series since 1955 to describe unexplained deteriorations in immunocompetent children during antituberculous therapy, but it has several limitations. Our inclusion criteria for culture-negative cases were quite strict. However, given that pediatric TB is often paucibacillary in nature, and that obtaining specimens for culture may be difficult, our clinical definition may have allowed for a few patients with other infectious processes, including bacterial or viral illnesses, to be included as active cases. We could not definitely rule out clinical failure in situations where the patient improved with multiple interventions, including addition of corticosteroids and antimicrobials to the regimen. However, despite sampling new lesions for culture in many instances, there were no findings of acquired drug resistance, and all patients received DOT and were eventually cured. Although data were recorded in a standardized way and management was by a consistent clinical and public health team using standardized protocols, the study was retrospective and instances of deterioration could have been missed. Referral bias may have led to more ill patients being included in the cohort. The sample size did not permit detailed analysis of risk factors for deterioration.
Our study, nonetheless, highlights deterioration during therapy as an important area in the treatment of children with TB. We found that these responses were common, could occur months into therapy, and could produce severe clinical consequences requiring immediate response. Clinicians managing pediatric TB should closely follow all children, especially those who are underweight or have extensive disease, for the occurrence of such reactions. As many of these episodes are probably immunopathologic in nature, corticosteroids should be considered in children who present with progression of disease or unexplained deterioration during therapy, including at sites remote from initial presentation, provided adequate antituberculous therapy is used concurrently. Further research in TB-endemic and low-incidence countries is needed to better define risk factors for deterioration and to improve recognition and management of these potentially life-threatening episodes.
The authors thank Marko Erak for his help with data collection, the staff of Toronto, Peel and York Public Health Units for their careful follow-up of patients during and after the study period, and Dr. Upton Allen and Dr. Anu Wadhwa for their review of the manuscript.
1. Cheng VC, Ho PL, Lee RA, et al.. Clinical spectrum of paradoxical deterioration during antituberculosis therapy in non-HIV-infected patients. Eur J Clin Microbiol Infect Dis. 2002;21:803–809.
2. Breen RA, Smith CJ, Bettinson H, et al.. Paradoxical reactions during tuberculosis treatment in patients with and without HIV co-infection. Thorax. 2004;59:704–707.
3. Hawkey CR, Yap T, Pereira J, et al.. Characterization and management of paradoxical upgrading reactions in HIV-uninfected patients with lymph node tuberculosis. Clin Infect Dis. 2005;40:1368–1371.
4. Chien JW, Johnson JL. Paradoxical reactions in HIV and pulmonary TB. Chest. 1998;114:933–936.
5. Narita M, Ashkin D, Hollender ES, et al.. Paradoxical worsening of tuberculosis following antiretroviral therapy in patients with AIDS. Am J Respir Crit Care Med. 1998;158:157–161.
6. Lawn SD, Bekker LG, Miller RF. Immune reconstitution disease associated with mycobacterial infections in HIV-infected individuals receiving antiretrovirals. Lancet Infect Dis. 2005;5:361–373.
7. Fishman JE, Saraf-Lavi E, Narita M, et al.. Pulmonary tuberculosis in AIDS patients: transient chest radiographic worsening after initiation of antiretroviral therapy. Am J Roentgenol. 2000;174:43–49.
8. Choremis CB, Padiatellis C, Zoumboulakis D, et al.. Transitory exacerbation of fever and roentgenographic findings during treatment of tuberculosis in children. Am Rev Tuberc. 1955;72:527–536.
9. Kumar R, Prakash M, Jha S. Paradoxical response to chemotherapy in neurotuberculosis. Pediatr Neurosurg. 2006;42:214–222.
10. Park JA, Park SS, Park SE. A paradoxical reaction during antituberculosis therapy for congenital tuberculosis. Int J Infect Dis. 2009;13:e279–e281.
11. Kitai I, Malloy P. A clinical approach to paediatric tuberculosis in Canada. Paediatr Child Health. 2003;8:162–170.
12. Kitai I, Malloy P, Long R. Pediatric
tuberculosis. In: Long R, Ellis E. eds. The Canadian Tuberculosis Standards. 6th ed. Ottawa. Ontario, Canada: Health Canada and the Canadian Lung Association;2007:308–319.
13. Tuberculosis Coalition for Technical Assistance. International standards for tuberculosis care (ISTC). The Hague, The Netherlands: Tuberculosis Coalition for Technical Assistance;2006.
14. Meintjes G, Lawn SD, Scano F, et al.. Tuberculosis-associated immune reconstitution inflammatory syndrome: case definitions for use in resource-limited settings. Lancet Infect Dis. 2008;8:516–523.
15. Loeffler AM. Pediatric
tuberculosis. Semin Respir Infect. 2003;18:272–291.
16. Cheng VC, Yam WC, Woo PC, et al.. Risk factors for development of paradoxical response during antituberculosis therapy in HIV-negative patients. Eur J Clin Microbiol Infect Dis. 2003;22:597–602.
17. Phongsamart W, Kitai I, Gardam M, et al.. A population-based study of tuberculosis in children and adolescents in Ontario. Pediatr Infect Dis J. 2009;28:416–419.
18. Kam A, Ford-Jones L, Malloy P, et al.. Active tuberculosis among adolescents in toronto, Canada: clinical features and delays in diagnosis. Pediatr Infect Dis J. 2007;26:355–356.
19. Cheng SL, Wang HC, Yang PC. Paradoxical response during anti-tuberculosis treatment in HIV-negative patients with pulmonary tuberculosis. Int J Tuberc Lung Dis. 2007;11:1290–1295.
20. Onal IK, Bayraktar Y, Unal S. Paradoxical deterioration during the course of antituberculous treatment. J Natl Med Assoc. 2006;98:954–955.
21. Meintjes G, Rabie H, Wilkinson RJ, et al.. Tuberculosis-associated immune reconstitution inflammatory syndrome and unmasking of tuberculosis by antiretroviral therapy. Clin Chest Med. 2009;30:797–810, x.
22. Murdoch DM, Venter WD, Feldman C, et al.. Incidence and risk factors for the immune reconstitution inflammatory syndrome in HIV patients in South Africa: a prospective study. AIDS. 2008;22:601–610.
23. Campbell IA, Dyson AJ. Lymph node tuberculosis: a comparison of various methods of treatment. Tubercle. 1977;58:171–179.
24. Pepper DJ, Marais S, Maartens G, et al.. Neurologic manifestations of paradoxical tuberculosis-associated immune reconstitution inflammatory syndrome: a case series. Clin Infect Dis. 2009;48:e96–e107.
25. Jain SK, Kwon P, Moss WJ. Management and outcomes of intracranial tuberculomas developing during antituberculous therapy: case report and review. Clin Pediatr (Phila). 2005;44:443–450.
26. Bukharie H. Paradoxical response to anti-tuberculous drugs: resolution with corticosteroid therapy. Scand J Infect Dis. 2000;32:96–97.
27. Garcia Vidal C, Rodriguez Fernandez S, Martinez Lacasa J, et al.. Paradoxical response to antituberculous therapy in infliximab-treated patients with disseminated tuberculosis. Clin Infect Dis. 2005;40:756–759.
28. Blackmore TK, Manning L, Taylor WJ, et al.. Therapeutic use of infliximab in tuberculosis to control severe paradoxical reaction of the brain and lymph nodes. Clin Infect Dis. 2008;47:e83–e85.
29. Wang ME, Castillo ME, Montano SM, et al.. Immune reconstitution inflammatory syndrome in human immunodeficiency virus-infected children in Peru. Pediatr Infect Dis J. 2009;28:900–903.
30. Markman M, Eagleton LE. Paradoxical clinical improvement and radiographic deterioration in anergic patients treated for far advanced tuberculosis. N Engl J Med. 1981;305:167.
31. Marais BJ, Gie RP, Schaaf HS, et al.. The natural history of childhood intra-thoracic tuberculosis: a critical review of literature from the pre-chemotherapy era. Int J Tuberc Lung Dis. 2004;8:392–402.
32. Nelson LJ, Schneider E, Wells CD, et al.. Epidemiology of childhood tuberculosis in the United States, 1993–2001: the need for continued vigilance. Pediatrics. 2004;114:333–341.