Paradoxical reactions in children with tuberculosis

Abstract

We sought to determine the incidence and clinical spectrum of the paradoxical reaction (PR) to treatment in 1000 children with tuberculosis (TB). Its incidence was 3.3%, presenting as mediastinal adenopathy (n = 12), central nervous system (CNS) tuberculoma (n = 8), increasing size of lymph nodes (n = 8) and serositis (n = 5). Symptoms included breathlessness (n = 7, 21.2%), neck swellings (n = 5, 15.2%), hemiplegia (n = 3, 9.1%), raised intracranial tension (n = 2, 6.1%) and fever or diabetes insipidus (n = 1, 3% each). The mean age of onset of PR was 3.5 months after the start of anti-TB medication. PR was seen predominantly in boys. Mediastinal lymphadenopathy was more common in children aged <2 years. Tuberculomas were more common in children aged 2–5 years. Nodal enlargement and serositis were seen equally in all age groups.

Keywords

Tuberculosis paradoxical reaction children serositis diabetes insipidus tuberculoma

Introduction

During the treatment of tuberculosis (TB), paradoxical reaction (PR) can occur. It is defined as the clinical or radiological worsening of the pre-existing tuberculous lesion or the development of a new lesion in a patient who initially improves with anti-tuberculosis therapy (ATT),¹ in the absence of disease relapse. It occurs in about 11–15% of patients with TB.² It is unpredictable in its timing (occurring anything from a few days to many months [range, 14–270 days] after the start of anti-tuberculosis chemotherapy) and in its duration and severity.¹ PR can occur in both HIV-uninfected and HIV-infected patients. It may manifest as recrudescence of fever, enlarging adenopathy, worsening pulmonary infiltration, pleural effusion, ascites, new or enlarging parenchymal CNS lesions, superior vena cava syndrome and involvement of bone, skin and soft tissue.^3–6 While these reactions are self-limiting, respiratory failure and death have been described, mainly with miliary and pulmonary disease.⁷ This study was undertaken to determine the incidence and clinical profile of children with PR.

Material and Methods

Our retrospective observational study was conducted after approval from the institution ethical committee. All children aged up to 15 years with TB and with PR who presented to the TB clinic during 2007–2012 were included. Records of all patients were evaluated; clinical history, examination findings and laboratory investigations were noted. Patients with drug-resistant TB on treatment with second-line therapy who had development of new lesions not related to treatment failure and were responding to the second-line therapy were also included. Patients who worsened owing to treatment failure, to non-compliance or to drug resistance, and HIV-infected patients and patients with congenital immunodeficiencies were excluded from the study.

PR was determined if there was clinical or radiological worsening of pre-existing the tuberculous lesion or the development of a new lesion in a patient who had initially improved with ATT and/or had become culture/smear-negative for TB or, in patients in whom cultures could not be done, when improvement was noted on continuing ATT. Worsening of radiological findings was defined when pleural or pericardial effusion appeared or increased, new parenchymal lesions developed with or without mediastinal lymphadenopathy, neuroimaging showed new lesions or increase in tuberculomata, or there was an increase in the size of existing lymph nodes.

Age, sex, onset of PR after treatment, duration of PR, types of PR and requirement of steroids for treatment of PR were noted and details of clinical history, examination, investigations and treatment were recorded. All patients underwent chest radiography and had Mantoux test, full blood count and erythrocyte sedimentation rates (ESR) done. Further investigations, such as culture/smear, were carried out as and when required. Cerebrospinal (CSF) PCR was done in patients who could afford it. Specific investigations such as biopsy, fine needle aspiration cytology, neuroimaging or chest imaging and ultrasound examination were done as required. Investigations were repeated whenever necessary. Patients were treated with standard ATT protocol as per RNTCP 2010 guidelines. No patients were under a directly observed treatment schedule (DOTS) and all were taking medicines as prescribed at home.

Descriptive data were analysed in proportion and percentage. Types of paradoxical reaction with various clinical parameters were analysed with χ² and Fisher’s exact test. P value < 0.05 was considered significant. Numerical data were analysed by statistical significance between the mean calculated by Student’s t-test. The risk of getting various types of paradoxical reactions with various factors such as age and gender was determined by odds ratio (OR). All data were statistically analysed using SPSS software version 18 and Primer.

Results

Thirty-three patients were diagnosed with PR out of a total of 1000 with tuberculosis. Table 1 depicts the demographic pattern of patients included in the study and the types of paradoxical reactions. Among the 19 patients who had a symptomatic PR, seven (21.2%) had breathlessness, five (15.2%) had new neck swellings, three (9.1%) had hemiplegia, two (6.1%) had ICT and one (3%) each had fever or diabetes insipidus. The mean age of presentation of PR was 4.3 years. The mean age of onset of PR was 3.5 months after the start of ATT. The duration of steroids required for PR was 3.0 months.

Table 1.

Demographic pattern of patients and type of paradoxical reactions.

Clinical features	Patients (n, %)
Age (years)
<2	13 (39.4%)
2–5	11 (33.3%)
>5	9 (27.3%)
Gender
Male	21 (63.6%)
Female	12 (33.4%)
Type of TB
Pulmonary TB with Mediastinal adenopathy	8 (24.2%)
Pulmonary TB	7 (21.2%)
Neurological TB	7 (21.2%)
Disseminated TB	6 (18.2%)
TB lymphadenopathy	3 (9.1%)
Abdominal TB	1 (3%)
Pleural effusion	1 (3%)
Drug resistance
MDR TB	4 (12.1%)
Paradoxical type
Mediastinal adenopathy	12 (36.4%)
Tuberculoma	8 (24.2%)
Increase in size of lymph node/abscess	4 (12.1%)
New cervical lymphadenopathy	4 (12.1%)
Pleural effusion	3 (9.1%)
Pericardial effusion	2 (6.1%)
Presentation
Symptomatic PR	19 (57.5%)
Steroids at the start of ATT
Yes	11 (33.3%)
PR while on steroids
Yes	7 (21.2%)
Treatment
Steroids for PR	26 (78.8%)
Required hospitalisation	8 (24.2%)

Factors associated with PR are depicted in Table 2. PR was seen predominantly in boys with OR for mediastinal adenopathy 1.96 (95% confidence interval [CI] = 0.38–9.9), for tuberculoma 2.7 (95% CI = 0.36–21), for new nodes or increase in size of pre-existing nodes 2.7 (95% CI = 0.36–21) and for serositis 16 (95% CI = 0.72–354). Mediastinal lymphadenopathy as PR was more common in children aged <2 years as compared to those aged 2–5 years (OR = 2.8; 95% CI = 0.53–14.7) and those aged >5 years (OR = 15.4; 95% CI = 1.47–160.9). OR on comparing mediastinal adenopathy between the age groups of 2–5 years and >5 years was 5.5 (95% Cl = 0.51–59). CNS tuberculomas were more common in children aged 2–5 years than in children aged <2 years (OR = 2; 95% CI = 0.36–24.9) and in those children aged >5 years (OR = 2; 95% CI = 0.36–24.9). New nodes or an increase in the size of nodes were seen equally seen in all age groups (OR = 0.55; 95% CI = 0.06–4.7). Similarly, serositis was seen equally in all age group (OR = 0.44; 95% CI = 0.03–5.58).

Table 2.

Clinical factors associated with various types of paradoxical reactions (PR).

	Mediastinal lymph nodes	Tuberculoma	New or increase in size of nodes	Pleural and pericardial effusion
Age (years)	2.1 ± 2.3	3.7 ± 2.1	6.6	7.1
Onset of PR after ATT started (months)	3.2 ± 2.9	6.1 ± 4.8	2.1	2.2
Initial dose of steroids for PR (mg/kg/d)	1.4 ± 1.0	1.8 ± 0.5	1.6	1
Duration of steroids (months)	2.6	3.8	2	3.7
ESR at start of ATT (mm/h)	41.4	15.3	40.8	17.6
Lymphocyte percentage at start of ATT	41.1	37.3	36.5	30
Clinical features	Mediastinal lymphadenopathy (n = 12)	Tuberculoma (n = 8)	New or increase in size of nodes (n = 8)	Pericardial and pleural effusion (n = 5)	P value
Gender
Male	7 (58.3%)	5 (62.5%)	5 (62.5%)	4 (80%)	0.886
Female	5 (41.6%)	3 (37.5%)	3 (37.5%)	1 (20%)
Age (years)
<2	7 (58.3%)	2 (25%)	2 (25%)	2 (40%)	0.225
2–5	4 (33.3%)	4 (50%)	3 (37.5%)	0
>5	1 (8.3%)	2 (25%)	3 (37.5%)	3 (60%)
Steroids for PR
Yes	9 (75%)	8 (100%)	4 (50%)	5 (100%)	0.055
No	3 (25%)	0	4 (50%)	0
PR while on steroids
Yes	1 (8.3%)	5 (62.5%)	0	1 (20%)	0.009
No	11 (91.6%)	3 (37.5%)	8 (100%)	4 (80%)

Discussion

In 1000 TB patients, paradoxical reactions were only seen in 33 and thus the sample size in our study was small. In adults, the incidence of PR has been reported as from 2.4% in pulmonary TB⁸ to 23% in lymph node TB,⁹ 23% in pleural TB¹⁰ and 17.2% in CNS tuberculosis.¹¹ PR has been reported in up to 35% of immunocompromised patients with HIV infection and in 2–10% in immunocompetent patients.^12,13 None of our patients had HIV and thus we cannot comment on the incidence of paradoxical reaction in HIV-positive patients as reported elsewhere.^8,14

It has been suggested that a higher bacillary load also predisposes patients to develop a PR.¹⁵ That the reaction can occur even after prolonged therapy suggests that the antigenic stimulus may be poorly cleared from disease sites. M. tuberculosis has a number of insoluble lipid-rich antigens in its cell wall that potentially invokes an inflammatory response, a mixed type 1 T helper cell (Th1) and type 2 T-helper cell (Th2) response¹⁵ and the activation of mononuclear phagocytes, and it may be that these components are responsible for the phenomenon of a deterioration in TB status.¹⁶ Thus, it would be expected that PR would be more frequent in adults owing to the multibacillary nature of the disease in adults and less frequent in children due to the paucibacillary nature of disease in children. However, in our study, paradoxical reaction was seen mostly in children aged <5 years. Apart from bacterial load, other body responses may also have a role in causing PRs which may lead to more young people being affected.^17,18

Lymph node enlargement either as mediastinal adenopathy or an increasing size of cervical nodes was a most common feature. More than half were symptomatic and approximately 25% had tachypnoea. Thus, children with pulmonary TB should be carefully monitored for respiratory distress and it should be noted that such clinical deterioration may suggest PR.

CNS tuberculoma was seen in more than one-quarter. Why the tuberculoma occurs on treatment is not properly understood. One possibility is that this occurs because of decreased penetration of anti-tubercular drugs into the brain. Restoration of the blood–brain barrier (BBB) with appropriate treatment is thought to result in reactivation of latent foci.¹¹ However, this hypothesis cannot explain the development or enlargement of intracranial tuberculomas which are treated with isoniazid and pyrazinamide, both of which freely cross the BBB in the absence of inflamed meninges. The reason for the occurrence of this response in only some cases and not all suggests that it depends on host immune responses, virulence of tubercle bacilli, antigen load and effective ATT. The improvement seen with steroid therapy for PR in neuro-tuberculosis may be due to a reduction in cerebral oedema and/or to a direct anti-inflammatory mechanism on the cerebral vasculature.¹⁹

PR were seen in approximately 20% with disseminated TB. This suggests that overall mycobacterial load can influence the development of PR. Campbell and Dyson²⁰ proposed that rapid killing of bacilli by effective ATT may release large amounts of tuberculoprotein and other cell wall products, generating a hypersensitivity reaction.

Five patients presented with serositis, of whom two had pericardial and three had pleural effusions. Pericardial effusion is a rare manifestation of PR. For pleural effusion, it is speculated that this results from the rupture of a subpleural pulmonary tubercular focus into the pleural space, which allows for tuberculoproteins to enter the pleural space and generate the hypersensitivity reaction responsible for a pleural effusion.²¹

The usefulness of corticosteroids in the treatment of PR is not well established and controversial.²² Corticosteroids are usually used for a severe systemic manifestation or a prolonged duration of PR. In our study, more than 75% of the patients with PR received steroids. Jung et al.¹⁰ found that patients treated with steroids had a shorter duration in improvement of pleural effusion compared with patients who did not. In our study, all patients with tuberculoma, pleural and pericardial effusions were treated with steroids. In fact, five (62.5%) with miliary TB and PR were already taking steroids and so required a higher dose. The use of steroids at the start of ATT needs to be assessed. Adjuvant therapy with corticosteroids, in conjunction with ATT, may be appropriate in particular forms of TB. Prospective controlled trials have shown a benefit in tuberculous meningitis, pericarditis and pleurisy. Although benefit has been shown in pleural disease, adjuvant therapy is not routinely required unless there are significant systemic symptoms of fever or a particularly large effusion. Though doses of 2 mg/kg/day of prednisolone have been routinely used, high doses of even 4 mg/kg/day have been tried in patients with CNS TB.^23,24

In conclusion, the occurrence of PR during ATT needs vigilance. Prompt recognition is essential to differentiate it from treatment failure or drug resistance. PR does not need change in ATT though sometimes the addition or dosage adjustment of steroids is needed to control symptoms.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Cheng

Lee

. Clinical spectrum of paradoxical deterioration during antituberculosis therapy in non-HIV-infected patients. Eur J Clin Microbiol Infect Dis 2002; 21: 803–809.

Cheng

Yam

Woo

. Risk factors for development of paradoxical response during antituberculosis therapy in HIV negative patients. Eur J Clin Microbiol Infect Dis 2003; 22: 597–602.

Corbella

Carratala

Rufi

. Unusual manifestations of miliary tuberculosis: cutaneous lesions, phalanx osteomyelitis, and paradoxical expansion of tenosynovitis. Clin Infect Dis 1993; 16: 179–180.

Afghani

Lieberman

. Paradoxical enlargement or development of intracranial tuberculomas during therapy: case report and review. Clin Infect Dis 1994; 19: 1092–1099.

Chien

Johnson

. Paradoxical reactions in HIV and pulmonary TB. Chest 1998; 114: 933–936.

Hill

Mateo

Hudak

. Transient exacerbation of tuberculous lymphadenitis during chemotherapy in patients with AIDS. Clin Infect Dis 1994; 19: 774–776.

Onwubalili

Scott

Smith

. Acute respiratory distress related to chemotherapy of advanced pulmonary tuberculosis: a study of two cases and review of the literature. Q J Med 1986; 59: 599–610.

Barss

. Unexpected deaths in pulmonary tuberculosis. Lancet 1983; i: 1437–1437.

Hawkey

Yap

Pereira

. Characterization and management of paradoxical upgrading reactions in HIV-uninfected patients with lymph node tuberculosis. Clin Infect Dis 2005; 40: 1368–1371.

10.

Jung

Shin

Kim

. Risk factors for development of paradoxical response during anti-tuberculosis treatment in HIV-negative patients with pleural tuberculosis. Tohoku J Exp Med 2011; 223: 199–204.

11.

Gupta

Bajaj

Khwaja

. Paradoxical response in patients with CNS tuberculosis. J Assoc Physicians India 2003; 51: 257–260.

12.

Breen

Smith

Bettison

. Paradoxical reactions during tuberculosis treatment in patients with and without HIV co-infection. Thorax 2004; 59: 704–707.

13.

Narita

Ashkin

Hollender

. Paradoxical worsening of tuberculosis following anti-retroviral therapy in patients with AIDS. Am J Respir Crit Care Med 1998; 158: 157–161.

14.

Shah

. Immune Reconstitution Syndrome in HIV-1 infected children - a study from India. Indian J Pediatr 2011; 78: 540–543.

15.

Moreno

Taverne

Mehlert

. Lipoarabinomannan from Mycobacterium tuberculosis induces the production of tumor necrosis factor from human and murine macrophages. Clin Exp Immunol 1989; 76: 240–245.

16.

Wallis

Amir-Tahmasseb

Ellner

. Induction of interleukin 1 and tumor necrosis factor by mycobacterial proteins: the monocyte Western blot. Proc Natl Acad Sci USA 1990; 87: 3348–3352.

17.

Song

Baek

Jung

. Paradoxical response developed during the antituberculous treatment intuberculous pleurisy. Tuberc Respir Dis 2008; 64: 427–432.

18.

Cho

Park

Kim

. Paradoxical responses in non-HIV-infected patients with peripheral lymph node tuberculosis. J Infect 2009; 59: 56–61.

19.

Dastur

Manghani

Udani

. Pathology and pathogenetic mechanisms in neurotuberculosis. Radiol Clin North Am 1995; 33: 733–752.

20.

Campbell

Dyson

. Lymph node tuberculosis: a comparison of various methods of treatment. Tubercle 1977; 58: 171–179.

21.

Al-Majed

. Study of paradoxical response to chemotherapy in tuberculous pleural effusion. Respir Med 1996; 90: 211–214.

22.

Dooley

Carpenter

Rademacher

. Adjunctive corticosteroid therapy for tuberculosis: a critical reappraisal of the literature. Clin Infect Dis 1997; 25: 872–887.

23.

Shah

Meshram

. High dose versus low dose steroids in children with tuberculous meningitis. J Clin Neurosci 2014; 21: 761–764.

24.

Evans

. The use of adjunctive corticosteroids in the treatment of pericardial, pleural and meningeal tuberculosis: do they improve outcome? Respir Med 2008; 102: 793–800.