Congenital tuberculosis is a rare disease that can mimic more common severe conditions affecting young infants. Diagnosis can be difficult, especially in nonendemic settings. We report a case of congenital tuberculosis complicated by a second rare condition, hemophagocytic lymphohistiocytosis (HLH) and review this unusual association.
A 3-week-old Australian-born male infant presented with fever, tachypnea, loose stools and poor weight gain. He had been born at term by vaginal delivery with a weight of 2900 g. On examination, there was no clear focus for the fever and he had no oxygen requirement. Initial investigations showed normal urine and cerebrospinal fluid findings, near-normal hematologic indices with slightly elevated white blood cell band forms (0.90 × 109/L), elevated C-reactive protein (84 mmol/L) and nonspecific patchy lung infiltrates in a chest radiograph.
His parents and a 6-year-old sibling had emigrated from India to Australia just before his conception. They were all in good health and had no personal history of tuberculosis or sick contacts. His mother had sought medical attention during pregnancy and been reassured in relation to a vaginal discharge late in the third trimester that persisted in the postpartum period.
After a full septic work-up, empiric intravenous antibiotics were commenced (cefotaxime 200 mg/kg/d 4 times a day, penicillin G 600,000 units/kg/d three times a day and gentamicin 5 mg/kg once daily). Blood, cerebrospinal fluid and urine cultures were sterile. He was discharged well at 72 hours receiving oral amoxicillin–clavulanate 45 mg/kg/d of the amoxicillin component, twice a day, for suspected lower respiratory tract infection.
At 6 weeks of age, he re-presented with fever, cough and respiratory distress with widespread coarse crepitations and patchy decreased air entry on auscultation. He was treated with intravenous penicillin G 400,000 units/kg/d four times a day and azithromycin 10 mg/kg once daily. A chest radiograph demonstrated bilateral patchy changes. Respiratory viral (influenza A and B, parainfluenza 1–3, respiratory syncytial virus and human metapneumovirus) and Bordetella pertussis by polymerase chain reaction (PCR) testing were not detected on nasopharyngeal aspirate.
After 24 hours, there was a marked clinical deterioration with the development of a new oxygen requirement, increasing fever, hepatosplenomegaly, rising C-reactive protein, abnormal liver function tests and pancytopenia. A bronchoscopy and bronchoalveolar lavage were done that revealed copious thick secretions and airway narrowing. After the bronchoscopy, he developed acute respiratory distress syndrome requiring intubation and high-frequency oscillatory ventilation, with progressive multiple organ dysfunction syndrome, including circulatory shock requiring epinephrine and norepinephrine infusions, and hepatic and renal failure. Antimicrobial treatment was broadened to include intravenous meropenem 80 mg/kg/d twice per day and trimethoprim–sulfamethoxazole 10 mg/kg/d twice per day (adjusted dosing for renal impairment). The family was counseled, and treatment limitations were discussed.
HLH was recognized in the constellation of fever, hepatosplenomegaly, pancytopenia (hemoglobin, 67 g/L; white cell count, 1.7 × 109/L; platelets, <10 × 109/L), elevated ferritin (5645 mcg/L; RR, 8–135), hypertriglyceridemia (3.3 mmol/L; reference range [RR], 0.9–2.0), hypofibrinogenemia (<0.5 g/L; RR, 0.8–3.8), elevated soluble CD25 (>50,000 pg/mL, RR 239–7887) and bone marrow hemophagocytosis.1 Extensive testing for congenital and acquired viral infections typically associated with HLH was noncontributory, and HIV serology was negative (Abbott CLIA Ag/Ab Combo Assay). Acid-fast bacilli were not detected by Ziehl–Neelsen smear microscopy of bronchoalveolar lavage fluid. However, Mycobacterium tuberculosis was detected in bronchoalveolar lavage fluid by PCR without detection of a mutation of the ribonucleic acid polymerase gene conferring rifampicin resistance (Xpert MTB/RIF, manufactured by Cepheid, Sunnyvale, CA). In addition, M. tuberculosis was isolated by culture of broncheoalveolar lavage fluid and tracheal aspirates after 3 weeks (BACTEC MGIT 960, manufactured by Becton Dickinson, Franklin Lakes, NJ). A QuantiFERON-TB Gold assay (Cellestis, Chadstone, Australia) was negative 1 month after initiation of tuberculosis treatment. Tuberculin skin testing was not done. Placental histology was reviewed and demonstrated no diagnostic features for tuberculosis. M. tuberculosis was not detected by PCR of fixed placental tissue.
Because of pharmacokinetic concerns in the context of critical illness and a desire to avoid further nephrotoxicity, initial antituberculous therapy comprised intravenous isoniazid 15 mg/kg/d, intravenous rifampicin 15 mg/kg/d, intravenous moxifloxacin 10 mg/kg/d and oral pyrazinamide 30 mg/kg/d, with intravenous dexamethasone 0.5 mg/6 h. A rapid and profound clinical and biochemical improvement was observed, enabling transfer to the ward from the intensive care unit within 96 hours. Central nervous system involvement was excluded by a repeat lumbar puncture and a magnetic resonance imaging scan. Abdominal ultrasound did not reveal a primary hepatic focus or significant intra-abdominal lymphadenopathy. Immune and genetic testing demonstrated no identifiable familial HLH genotype.
Drug-induced hepatotoxicity occurred in the first week of treatment with peak alanine transaminase of 504 units/mL (reference range, 10–40 units/mL), γ-glutamyl transpeptidase of 629 units/mL (RR, 0–225 units/mL) and bilirubin of 160 mmol/L (RR 0–10 mmol/L). Isoniazid and pyrazinamide were temporarily withheld, with substitution of intravenous amikacin 18 mg/kg/d and oral ethambutol 20 mg/kg/d. Liver function initially improved and then worsened again on the reintroduction of isoniazid, normalizing when it was ceased. Pyrazinamide was reintroduced later without event, and intravenous amikacin was ceased. Drug susceptibility testing subsequently demonstrated isoniazid resistance only.
The parents, a 6-year-old sibling and maternal grandparents visiting from India had all previously received Bacille Calmette Guerin vaccination, and none had clinical or radiological features of pulmonary tuberculosis. With the history of a vaginal discharge during pregnancy, the infant’s mother was referred for investigation for endometrial tuberculosis. The diagnosis of congenital tuberculosis was confirmed by maternal endometrial biopsy (see Fig., Supplemental Content 1, https://links.lww.com/INF/C284) with histological evidence of granulomatous endometritis with epithelioid histiocytes, lymphocytes and multinucleated giant cells, and subsequent isolation of M. tuberculosis by culture with isoniazid resistance. The mother subsequently completed 9 months of tuberculosis treatment.
After 2 months, the infant was discharged with nasogastric nutritional support, tapering corticosteroids (ceased completely after 10 weeks) and continuing tuberculosis treatment with oral moxifloxacin, rifampicin and pyrazinamide (doses as above). Pyrazinamide was ceased 3 months after initiation of tuberculosis treatment. Moxifloxacin and rifampicin were ceased at 12-month follow-up, with the infant thriving and completely recovered.
This report highlights a case of congenital tuberculosis triggering the development of HLH. Timely diagnosis and specific therapy led to a favorable outcome for the patient and also for his mother. Congenital tuberculosis is a rare diagnosis in countries where tuberculosis prevalence is low, and HLH is also very uncommon.
HLH is a hyperinflammatory disorder resulting from immune dysfunction reflecting either primary immune deficiency or acquired failure of normal immune homeostasis.2 HLH may be triggered by a relatively minor insult where there is an underlying genetic predisposition, or, in the absence of such a genotype, it occurs at a relatively higher inflammatory threshold triggered by severe infection such as in this case report, or a malignancy or autoimmune disorder.3 A search of the English language literature identified a further 13 reports of tuberculosis complicated by HLH in children or adolescents (Table 1).4–16 The age of affected patients at presentation ranged from 2 weeks to 17 years, and with the inclusion of the current case, there were 8 patients aged 2 months or less. Three patients died, even despite combined immunomodulatory therapy and antituberculous treatment in two. In 1 fatal case, from a nonendemic setting (US) and despite extensive investigation, the diagnosis of tuberculosis remained unknown until an autopsy was performed. Of 11 surviving patients, 3 received only antituberculous treatment and 8 received both antituberculous treatment and immune therapies (4 had corticosteroids, 5 had intravenous immunoglobulin, 1 cyclosporine and 1 etoposide). In 10 of 14 cases, there was evidence of disseminated tuberculosis, including in 6 of 8 patients aged 2 months or less, with pulmonary disease in the remaining 4 cases. All patients presented with fever, 13 of 14 had organomegaly and 13 of 14 had thrombocytopenia and/or pancytopenia. Diagnostic samples for tuberculosis included bronchoalveolar lavage, tracheal aspirates, gastric aspirates, lymph node biopsy, liver biopsy, lung biopsy, bone marrow and blood culture. A 2006 review, including 36 adult and pediatric cases, found approximately half of adult cases were associated with comorbidities, including end-stage renal failure, AIDS and solid organ transplant.17 The neonatal period and very early infancy may likewise be considered a period of relative immune compromise.
Recognition of uncommon critical presentations of uncommon conditions requires that clinicians avoid the pitfall of premature diagnostic closure and remain vigilant for the failure of standard empirical diagnostic and management approaches. The patient history should be revisited and expanded, and an expedited and aggressive search for a definitive diagnosis should be conducted, which may require invasive second-line investigations, such as bronchoscopy and bone marrow examination, in this case.
We acknowledge Stephanie A. Brumby (Pharmacy Department, The Royal Children’s Hospital Melbourne, VIC, Australia), Jackie Collett (Department of Pathology, Royal Women’s Hospital, Melbourne, VIC, Australia), Renata Kukuruzovic (Department of General Medicine, The Royal Children’s Hospital Melbourne, VIC, Australia), Stephanie Richards (Department of Allergy and Immunology, The Royal Children’s Hospital Melbourne, VIC, Australia), and Alan Street (Victorian Infectious Diseases Unit, The Royal Melbourne Hospital, Parkville, VIC, Australia).
1. Henter JI, Horne A, Aricó M, et al. HLH-2004: Diagnostic and therapeutic guidelines for hemophagocytic lymphohistiocytosis
. Pediatr Blood Cancer. 2007;48:124–131
2. Usmani GN, Woda BA, Newburger PE.. Advances in understanding the pathogenesis of HLH. Br J Haematol. 2013;161:609–622
3. Risma K, Jordan MB.. Hemophagocytic lymphohistiocytosis
: updates and evolving concepts. Curr Opin Pediatr. 2012;24:9–15
4. Dey A, Shah I, Paikrao P, et al. Tuberculosis with hemophagocytic lymphohistiocytosis
in an infant. Indian J Pediatr. 2014;81:214–215
5. Gupta AP, Parate SN, Bobhate SKAnupriya. . Hemophagocytic syndrome: a cause for fatal outcome in tuberculosis. Indian J Pathol Microbiol. 2009;52:260–262
6. Verma T, Aggarwal S.. Childhood tuberculosis presenting with haemophagocytic syndrome. Indian J Hematol Blood Transfus. 2012;28:178–180
7. Maheshwari P, Chhabra R, Yadav P.. Perinatal tuberculosis associated hemophagocytic lymphohistiocytosis
. Indian J Pediatr. 2012;79:1228–1229
8. Deshpande A, Nayar PS, Pradhan AM, et al. Miliary tuberculosis with hemophagocytosis in a two months old infant. Indian J Hematol Blood Transfus. 2010;26:115–117
9. Balasubramanian S, Kaarthigeyan K, Aparna V, et al. Tuberculosis associated hemophagocytic syndrome in infancy. Indian Pediatr. 2008;45:593–595
10. Tavil B, Caliskan U, Unal S, et al. Pulmonary tuberculosis presenting with pancytopaenia, haemophagocytosis and foamy histiocytes in an infant. Int J Tuberc Lung Dis. 2007;11:931–932
11. Okascharoen C, Nuntnarumit P, Sirinavin S.. Neonatal tuberculosis associated with shock, disseminated intravascular coagulation, hemophagocytic syndrome, and hypercalcemia: a case report. J Perinatol. 2003;23:79–81
12. Akinbami LJ, Selby DM, Slonim AD.. Hepatosplenomegaly and pulmonary infiltrates in an infant. J Pediatr. 2001;139:124–129
13. Shaw PH, Brown D, Shulman ST.. Tuberculosis-associated hemophagocytic syndrome in an infant. Pediatr Infect Dis J. 2000;19:475–477
14. Chen CH, Fang YH, Chiang PM, et al. Disseminated tuberculosis presenting as multiple hepatosplenic microabscesses and pancytopenia in a teenage boy. J Formos Med Assoc. 2004;103:939–942
15. Dilber E, Erduran E, Kalyoncu M, et al. Hemophagocytic syndrome as an initial presentation of miliary tuberculosis without pulmonary findings. Scand J Infect Dis. 2002;34:689–692
16. Monier B, Fauroux B, Chevalier JY, et al. Miliary tuberculosis with acute respiratory failure and histiocytic hemophagocytosis. Successful treatment with extracorporeal lung support and epipodophyllotoxin VP 16-213. Acta Paediatr. 1992;81:725–727
17. Brastianos PK, Swanson JW, Torbenson M, et al. Tuberculosis-associated haemophagocytic syndrome. Lancet Infect Dis. 2006;6:447–454