Case 2 -

Congenital (Perinatal) Intrauterine Tuberculosis (TB) with Emphasis on Placental Findings Carlos R. Abramowsky Emory University Egleston Children's Hospital Atlanta, GA

Click on slide thumbnail images for an enlarged view.

If you have any difficulties viewing these slides, email the webmaster.

Clinical Summary An adequate for gestation newborn female was delivered at term by elective cesarean section, with good Apgar scores. The infant's initial evaluation was unremarkable. The mother was a healthy 31 year old, well to do African from Liberia and the pregnancy was uncomplicated. However, intractable postpartum bleeding led to a hysterectomy with a diagnosis of retained placenta. The placenta and uterus were examined in the pathology department with extensive sampling of the placenta. A diagnosis of placenta accreta was rendered. The newborn developed fever and leucocytosis in the first week of life and she was treated for bacterial sepsis although all cultures were eventually negative. Her physical examination was unremarkable except for slight hepatomegaly. A TORCH and other viral and malaria studies were negative, but serum IgM was 250 mg/dL (normal: up to 25mg/dL), consistent with a congenital infection. Head imaging was negative but an eye exam showed chorioretinitis. In the first 3 weeks she appeared mildly ill with lethargy and poor feeding. At day 20 postpartum, the mother developed a respiratory infection with pleural effusion and her clinical workup led to the diagnosis of pulmonary tuberculosis. On this basis, a similar diagnosis was entertained in the newborn when she developed classic miliary disease in her lungs (Figure 1) and hepato-splenomegaly. A PPD skin test showed 6mm of induration but a gastric lavage was positive for Mycobacterium tuberculosis. A review of the many samples of placenta which had been obtained to address the issue of postpartum hemorrhage was undertaken. A single focus of acute villitis and intervillitis with a predominant neutrophilic response and no evidence of granulomatous inflammation was identified (Figures 2,3). An Auramine-O fluorescent stain for acid fast bacteria was positive in this lesion (Figure 4). Anti-tuberculosis therapy was started with INH and Streptomycin. Two days later the child became critically ill with hypoxemia and DIC, worsening liver chemistries and respiratory failure requiring ventilation. Steroids and Pyrazinamide were added with good improvement of her respiratory symptoms. However, after several weeks her liver disease worsened with elevated bilirubin and transaminases and coagulation abnormalities. All these problems were treated with supportive measures with good results except for continuing liver chemistry abnormalities. A liver biopsy revealed multiple non-caseating granulomatous hepatitis with positive acid fast bacteria on microscopic examination and culture (Figures 5-7), despite several weeks of anti-TB therapy. Yet, the patient continued to improve and was discharged at 2 months of age. She remained with mild liver function abnormalities, mild hypersplenism and poor weight gain. At 9 months of age she had improved substantially, gained weight and towards her second year of life had recovered completely with no residual deficits. Anti-TB therapy was suspended at one year of age. Case 2 - Figure 1 Chest Xray of the newborn, shows diffuse densities bilaterally, consistent with a miliary infection. Case 2 - Figure 2 These low and higher magnification views, show the focus of villitis and intervillitis with a primarily neutrophilic inflammatory response (H & E stain). Case 2 - Figure 3 These low and higher magnification views, show the focus of villitis and intervillitis with a primarily neutrophilic inflammatory response (H & E stain). Case 2 - Figure 4 A sequential section of the same focus shown in figures 2 and 3 shows several acid fast bacilli consistent with mycobacteria, using an Auramine-O fluorescent method. Case 2 - Figure 5 At several weeks of age, and with persistent abnormal liver function tests, a liver biopsy of the infant showed non-necrotizing granulomatous hepatitis. Case 2 - Figure 6 At several weeks of age, and with persistent abnormal liver function tests, a liver biopsy of the infant showed non-necrotizing granulomatous hepatitis. Case 2 - Figure 7 An Auramine-O fluorescent stain on the liver biopsy also showed acid fast bacilli which were confirmed to be M.tuberculosis in liver biopsy cultures. Case 2 - Figure 8 This view is from the decidual layers of a placenta from a patient different from our index case, who had chronic tuberculosis and tuberculous meningitis at the time of delivery. The photomicrograph shows necrotizing granulomatous inflammation in the decidua, and the accompanying acid fast stain shows an occasional mycobacterial organism (arrows). There was no evidence of TB infection in the newborn who was treated immediately postnatally. Case 2 - Figure 9 This view is from the decidual layers of a placenta from a patient different from our index case, who had chronic tuberculosis and tuberculous meningitis at the time of delivery. The photomicrograph shows necrotizing granulomatous inflammation in the decidua, and the accompanying acid fast stain shows an occasional mycobacterial organism (arrows). There was no evidence of TB infection in the newborn who was treated immediately postnatally. Pathologic Diagnoses -Pathologic and clinical findings diagnostic for perinatal (congenital-transplacental) tuberculosis: a-Third trimester placenta with focal acute villitis and intervillitis showing a predominant neutrophilic inflammatory reaction and no evidence of a granulomatous component. Auramine-O fluorescent stain for acid fast bacteria was positive for bacteria consistent with M. tuberculosis. b-Newborn liver (needle biopsy): Mycobacterial granulomatous hepatitis. The mycobacteria were revealed with the Auramine-O technique and species identification with liver biopsy culture. c-Third trimester placenta with deficient decidual layer and direct implantation on myometrium consistent with placenta accreta (images not shown). e- Uterus with hypoplasia of basal endometrial layer and involuting placental site. Cervix: Chronic cervicitis and nabothian cysts. There is no evidence of granulomatous or unusual inflammation in the specimen (images not shown) Diagnoses: Perinatal tuberculosis with miliary disease involving the liver and lungs of a newborn. Focal acute (non-granulomatous) placental villitis and intervillitis with acid fast mycobacteria, consistent with in-utero transplacental infection. Discussion Tuberculosis (TB) is one of the most widely disseminated diseases with some 2 billion patients worldwide and 9 million new cases reported in 2005. Despite those numbers, it appears that only about 400 cases of congenital or perinatal TB have been reported throughout the last century, and very few of these have had adequate placental evaluations [1]. Perinatal TB had been rarely reported in developed countries until more recently as immigration from the developing world has increased [2]. The AIDS epidemic has also caused a spike in maternal TB infections and thus its perinatal version [1]. In a South African study (cited in 3), of 107 HIV + mothers with TB during pregnancy, mycobacteria were isolated in 15% of the newborns at 1-3 weeks postnatally. A fetus-newborn may be infected with TB: by transplacental or hematogenous transmission, the hallmark of which is hepatic lesions derived from mycobacteria in umbilical vein blood. By inhalation of infected amniotic fluid, usually manifested by pulmonary infiltrates. By infection from the maternal genital tract at delivery. By postnatal infection from contact with the mother, other family members or healthcare providers. The criteria for the diagnosis of perinatal TB (previously: congenital TB) were first defined in 1935 and modified in 1994 [4]: The newborn shows lesions in the first week of life there is evidence of a primary hepatic complex or biopsy evidence of hepatic granulomas evidence of tuberculous infection of the placenta or maternal genital tract exclusion of postnatal tuberculous infection by family or hospital workers. Infection of the fetus or newborn is even less frequent if the mother is known to have had TB and especially if she has received treatment. The untreated chronic TB patients have low fertility associated with their weakened state, and their lesions consist of chronic necrotizing granulomatous inflammation with variable fibrosis, low levels of tissue mycobacteria and negligible or no organisms in the blood and genital secretions. Therefore, many cases of perinatal TB seem to be associated with a concomitant acute or new infection in the mother. This makes sense, because in a new and recent infection, mycobacteremia is more likely, which is a necessary requirement for hematogenous transplacental transmission. Our index case reflects the many reports in which the mother was apparently healthy, without family history of TB or other infectious problems. Indeed, it was her postnatal acute onset of lung disease and subsequent TB diagnosis that led to the eventual diagnosis in the child who had remained a clinical puzzle. Also, as in our index case, these perinatally infected newborns do not manifest signs or symptoms for several days to weeks. At first they appear to be septic with fever, leucocytosis (mostly neutrophilic), lethargy etc. The use of standard antibiotics for bacterial sepsis have no effect and other signs start to appear such as hepato-splenomegaly and pulmonary infiltrates. For several reasons, the skin test may not be positive in the newborn for weeks or months, but gastric lavage provides a good yield for diagnosis [4]. Placenta: The delay in clinical onset of disease and diagnosis of the newborn (and often the mother) is the most likely explanation for the scanty reports of placental pathology in congenital TB, as this organ would likely have been discarded when the diagnosis is finally considered several weeks postpartum. In our index case, the placenta (and uterus) had been evaluated and extensively sampled serendipitously because of the post-partum hemorrhagic complications. Placental findings in perinatal tuberculosis include: chronic necrotizing granulomas in the decidua or endometrium, if the uterus is also examined. These lesions are rare in this location, and tend to reflect a more established or chronic maternal infection (Figures 8,9). If sufficiently active and productive of mycobacteria, they are a potential source of fetal infection in the birth canal. A necrotizing granulomatous inflammation was not seen in the many sections of placenta or the uterine lining of our index case. Placental inflammation: inkeeping with other transplacental infections, placental villitis often with intervillitis, is the hallmark lesion. While granulomatous inflammation is to be expected and has been reported in cases of tuberculosis, a non-specific acute neutrophilic (non-granulomatous) reaction was seen in our index case, in another case in our files, and in a few published reports [4, 5]. In our index case, the focus of acute villitis-intervillitis (figures 2,3) was seen in only one of many blocks of placenta. This focality of placental lesions has been mentioned in older reports and may be another reason for the negative placental findings in some of them if not enough blocks are examined. On the other hand, some of the older reports, including the first confirmed case of congenital tuberculosis in 1891 [6] and early 20th century (cited in [7, 8] ) indicate the presence of acid fast bacteria within the placenta and fetus, without an inflammatory reaction. In those pre-antibiotic days, many patients suffered fulminant TB infections causing rapid death of the fetus and mother; under those conditions an inflammatory reaction may not have had time to evolve. Why a non-granulomatous, acute neutrophilic inflammatory response? It is very likely that an acute, neutrophilic reaction to TB as seen in the index case and in others in our experience, represents a first contact response to the mycobacteria. This is a rarely witnessed morphological event in the usual setting of tuberculosis of the lung. However it has been described in some of the pathology textbooks throughout the 20th century. For example, Forbus, in his 1943 "Reaction to Injury" [9] shows a photomicrograph of a hepatic neutrophilic abcess from a patient with tuberculous miliary disease (Figure 373, page 582). In the same text, experimental inoculation of mycobacteria in rabbits results in: "The bacilli (in the lung) are quickly phagocytosed in the capillaries by the polymorphonuclear leucocytes. Within from 18-24 hours, the mononuclear wandering cells, the macrophages, accumulate within the capillaries, and from that time on they play the most conspicuous role..." (page 583). In: Pathology of the Human Placenta [10] the authors state that: "In other cases we have identified acute granulomas (consisting mostly of neutrophils) adjacent to villi (sic)..." In another report of two patients with perinatal tuberculosis, the authors found no lesions in the placenta but state that the lesions in one of the newborns at autopsy, were non-granulomatous [11]. It is important to mention that a neutrophilic inflammatory response in the villous placenta includes in the differential: a bacterial infection caused by Listeria monocytogenes, pyogenic Staphylococci, Streptococci or Pseudomonas, and a rare case of transplacental Herpes simplex virus infection, etc. An acute, neutrophilic inflammatory reaction in tuberculosis is likely due to a response mediated by the Innate Immune System. The role of the innate immune system in many conditions such as asthma, atopy, and other inflammatory disorders has been recently emphasized [12, 13]. Innate immunity has served the animal kingdom well as it is the only system in invertebrates who certainly have survived successfuly. The acquired or adaptive immune system evolved with vertebrates. In the case of TB, the adaptive immune system is predominantly responsible for development of granulomas and caseating necrosis, both hallmarks of the pathology of tuberculosis [14]. In patients with a defective immune system, a granulomatous reaction either does not develop, or is poorly formed. An autopsy study of TB infection in 20 AIDS patients with a median CD4 count of 0.065x109/L, showed acute, neutrophilic tuberculous meningitis and poorly formed granulomas elsewhere [15]. Patients with genetic imunodeficiencies also show diffuse histiocytic or poor granulomatous responses when inoculated with live Mycobacterium bovis in BCG vaccines [16, 17]. The innate immune system [12] has several components: the physical barrier system: skin, intestinal mucosa, gastric acid, lyzozyme in tears and saliva. The innate system proper, relies on a limited repertoire of receptors in neutrophils, macrophages and dendritic cells among other cells and humoral factors which target conserved cell surface molecules that are common to many pathogens. Among many cytokines and chemokines, some of the major players include: Toll-like receptors (TLRs) and Nucleotide oligomerization domain receptors (NOD or NLRs). This system serves as a rapid first line of defense against infections while the more delayed but specific adaptive immune response is turned on, also with the help of the innate system. The neutrophilic reaction to TB seen in our index and other cases seems to reflect an innate immune response of the mother in her first encounter with this pathogen, but also of the fetus in his first encounter with any pathogen. The fetus lives in an ostensibly germ-free environment in which his responses are TH2 dominant and geared to develop tolerance to newly formed antigens. Upon birth, or if congenitally infected, the fetus is suddenly exposed to foreigh antigens which he deals with by using his innate immune system and a rapidly turned on adaptive immunity [18, 19]. A neutrophilic and macrophagic response in the innate immunity phases of TB infections is documented in several studies in humans as mentioned and in experimental animals [20, 21, 22, 23, 24]. However, the clinical and experimental literature shows conflicting results as to whether the innate neutrophilic response is actually protective or whether it only contributes to tissue injury. The findings in our index case would suggest that that first encounter with the mycobacteria, presumably in that placental focus, was not a protective one. References Whittaker E. et al. Perinatal tuberculosis. New challenges in the diagnosis and treatment of tuberculosis in infants and the newborn. Early Human Development 84:795, 2008 Connelly Smith K. Congenital tuberculosis: a rare manifestation of a common infection. Current opinion in infectious diseases 15:269, 2002 Pillay T et al. Perinatal tuberculosis and HIV-1: considerations for resource-limited settings. Lancet Infect Dis 4:155, 2004 Cantwell MF et al. Congenital tuberculosis. NEJM 330:1051, 1994 Naranbhai RC et al. Congenital tuberculosis localized to the ear. Arch Dis Childh Schmorl, Birch-Hirschfeld. Uebergang von tuberkelbacillen aus dem mutterlichen blut auf die frucht. Beitrage zur pathologischen anatomie und zur allgemeinen pathologie 9: 428, 1891 Hughesdon MR. Congenital tuberculosis. Arch Dis Child 21:121, 1946 Grenville-Mathers R. et al. Tuberculous primary infection in pregnancy and its relation to congenital tuberculosis. Tubercle, Lond 41:181, 1960. Forbus WD. Chapter 29: Reaction to injury by Mycobacterium tuberculosis through granulomatous inflammation and antibody formation-the development of tuberculous infection in the experimental animal and in man. In: Reaction to Injury, The Williams and Wilkins Company, 1943 Benirschke K. et al. Chapter 20, p. 683: Infectious Diseases. In: Pathology of the Human Placenta, 5th Edition, Springer, 2006 Machin GA. et al. Perinatally acquired neonatal tuberculosis: report of two cases. Pediatric Pathology 12:707, 1992 Turvey SE. et al. Innate Immunity. J.Allergy Clin Immunol. 125:S24, 2010 Innate immunity-Special section. Science 327: 283, Jan 15,2010 Saunders BM et al. Life and death in the granuloma: immunopathology of tuberculosis. Immunology and cell biology 85:103, 2007 Smith MB et al. Generalized tuberculosis in the acquired immune deficiency syndrome. Arch Pathol Lab Med 124:1267, 2000 Abramowsky C et al. Disseminated bacillus Calmette-Guerin infections in patients with primary immunodeficiencies. Am J. Clin Pathol 100:52, 1993. Jean-François E. et al. Correlation of granuloma structure with clinical outcome defines two types of idiopathic disseminated BCG infection. Journal of Pathology 181:25, 1997 Levy O. Innate immunity of the newborn: basic mechanisms and clinical correlates. Nature reviews-immunology 7:379, 2007 Fleer A et al. Innate immunity: toll-like receptors and some more. Neonatology 92: 145, 2007 Korbel DS et al. Innate immunity in tuberculosis: myths and truth. Microbes and infection 10:995, 2008 Martineau AR et al. Neutrophil-mediated innate immune resistance to mycobacteria. JCI 117:1988, 2007 Gonçalves A-S et al. The involvement of the chemokine receptor CXCR2 in neutrophil recruitment in LPS-induced inflammation and in Mycobacterium avium infection. Scand J. Immunol 55:585, 2002 Kisich KO et al. Tumor necrosis factor alpha stimulates killing of Mycobacterium tuberculosis by human neutrophils. Infection and Immunity 70:4591, 2002 Eruslanov EB et al. Neutrophil responses to Mycobacterium tuberculosis infection in genetically susceptible and resistant mice. Infection and Immunity 73:1744, 2005.