This 2-year-old girl was admitted to the hospital for evaluation of an abnormal gait. She had been in her usual state of health until 1 month prior to this visit when she fell off a chair and landed on her back. Since that time, she had had increasing pain while walking. She was seen 2 weeks prior to this visit and had normal lower extremity x-rays in an outside emergency room. During the past week, she had been walking on her tiptoes, with apparent pain. There was no history of fever, vomiting, diarrhea, rashes, or other symptoms. Her medical history, family history, and birth history were unremarkable.
On physical examination, she was alert and healthy appearing. Her temperature was 37.2 degrees C, pulse 88, respiratory rate 20, and blood pressure 110/55. Weight was in the 10th percentile and height in the 5th percentile. HEENT exam was unremarkable. There was no significant adenopathy. Neck was supple. Lungs were clear. Cardiac examination was normal. S1 and S2 were normal without murmurs or rubs. The abdomen was soft and nontender, without masses or organomegaly.
Her back was straight; there was no kyphosis. There was no vertebral tenderness. Muscle strength in all the extremities was normal, save for the right lower extremity. She had 4/5 strength in the knee and ankle. Hip girdle strength was normal. Rectal tone was normal. Deep tendon reflexes were reportedly normal, with downgoing toes bilaterally.
Robert Listernick, MD, moderator: Can we make sense of her physical exam?
Douglas Nordli, MD, pediatric neurologist: First, a history of trauma is so ubiquitous in childhood that it has to be taken with a large grain of salt before attributing it as the cause of this girl's problems. The striking part of the history is the toewalking, which implies spasticity. However, her documented examination reveals normal deep tendon jerks and normal Babinski reflexes. This is difficult to reconcile.
Dr. Listernick: What about her motor exam?
Dr. Nordli: The description of her strength is inadequate. You rarely get a good confrontational exam from a 2-year-old. To assess hip and girdle strength, one should just watch the child walk around and climb onto things. This will give a much better sense of true leg weakness. Her exam suggests sacral sparing, as rectal tone was normal, but there's no history given as to whether there's any bowel or bladder involvement. At the moment, we don't know much except that we should be suspicious of a spastic paraparesis.
Dr. Listernick: My sense is that the physicians who first cared for her were concerned about the possibility of an epidural hematoma secondary to the trauma.
Dr. Nordli: Perhaps, but that has to be an extremely rare condition unless she were to have a bleeding disorder. Despite the ambiguous neurologic exam, we should be concerned about a spinal cord mass.
Dr. Listernick: The obvious first step is some form of imaging. Should we obtain plain films of the spine first?
John Curran, MD, pediatric neuroradiologist: It certainly would have been reasonable to get plain films. However, regardless of the result, further imaging would have been necessary. Negative plain films wouldn't have obviated the need for magnetic resonance imaging (MRI). Even if films had been positive, you would have wanted to better define the lesion that was found.
Dr. Listernick: I just wanted to make the point that sometimes we forget to do the simple things. Several years ago, we discussed a case of fever of unknown origin in a child who eventually was found to have disseminated histoplasmosis. First, he had a gallium scan, which was positive in the lung parenchyma, which led to the performance of a computerized tomography (CT) scan, which defined the parenchymal disease. However, he had never had a simple chest x-ray, which would have led immediately to the correct diagnosis without the need for all the expensive testing. So, what's the correct imaging procedure here?
Dr. Curran: If you're worried about a spinal cord or disc problem, the correct test would be a MRI.
MRI best defines a bone marrow process such as one would see in osteomyelitis. A CT scan would better delineate a cortical bony process.
An MRI scan was performed. The L5-S1 disc space is abnormal, with a soft tissue mass that is expanding out of it both anteriorly and posteriorly. The S1 vertebral body is partially destroyed. The mass has areas of irregular enhancement surrounding areas of nonenhancement. Its appearance is that of either a tumor with a necrotic center or an abscess; it's difficult to distinguish between these two possibilities.
Dr. Listernick: What kind of tumor might this be?
Reggie Duerst, MD, pediatric oncologist: I would suspect that the absence of fever makes abscess less likely. By far and away, the most common tumor of young children to present with spinal cord involvement would be a neuroblastoma. It often has calcifications within its center due to necrosis as the tumor outgrows its blood supply. Less likely would be either Ewing's sarcoma or a primitive neuroectodermal tumor.
A solitary eosinophihc granuloma might cause vertebral body collapse, but this amount of bony destruction and necrosis would be unusual.
Dr. Listernick: What about abscess?
Ben Katz, MD, pediatric infectious disease physician: When I first heard about the case, I thought that infection was unlikely due to the absence of fever. However, if you were to consider an infection, it would probably be a subacute or chronic infection, given the prolonged time course and the lack of inflammatory symptoms. The list would include such infections as tuberculosis (TB) or fungal diseases, both distinctly uncommon.
Dr. Listernick: You don't think that this could be run-of-the-mill staphylococcal osteomyelitis that started with the fall and back trauma?
Dr. Katz: I don't think that the MRI, with its large soft tissue mass and bony destruction, looks like an involucrum (formation of new bone around an area of dead bone).
Ellen Chadwick, MD, pediatric infectious disease physician: I agree with Ben but would point out that fever is less likely to occur in a chronic bone infection.
Dr. Listernick: What's the next step?
Dr. Katz: As my oncologist colleague often says, "Tissue is the issue." The question is how one obtains this tissue. We need enough of the specimen for culture, Gram stain, special stains, and histology. I'd have to talk with the neurosurgeons and the interventional radiologists about which was the safest procedure that could maximize the yield.
Dr. Listernick: The first procedure performed was a fine needle aspirate, which did not yield any useful information.
Reggie Duerst, MD, pediatric oncologist: In general, a fine needle aspiration is a reasonable first step, particularly if you expect to find pus. If it's a necrotic tumor, I think that you're unlikely to make a definitive diagnosis with a fine needle. If it's an infection, I would think that she would need an extensive debridement to prevent cord compression and to control the infection. Therefore, regardless of the result of the aspiration, it seems that a bigger procedure would be necessary.
Dr. Nordli: The neurosurgeons aren't here to comment, but they must have been fairly unimpressed with the physical examination and the risk of irrevocable cord damage to have proceeded in this cautious, step-wise fashion.
Elaine Morgan, MD, pediatric oncologist: I'd offer a different viewpoint For example, if this child had metastatic neuroblastoma, removal of the lesion would not be indicated and could be associated with extensive morbidity. A good core biopsy may be all that is necessary.
Dr. Listernick: Is there a problem with giving this child corticosteroids at presentation to reduce cord edema?
Dr. Morgan: If she had lymphoma, an unlikely diagnosis at this age and location, steroid administration might make it difficult to assess the soft tissue component adequately. The bony component will not disappear as readily. Steroid treatment should not significantly affect any other soft tissue tumor.
Dr. Listernick: Moving on, there's some important epidemiologic information that I left out for the sake of the discussion. After multiple historians asked this family multiple questions on multiple occasions, one of the infectious disease physicians asked specifically about a family history of tuberculosis (TB). The father finally said that he was taking four medications to treat TB. Eventually (during the first 24 hours in the hospital), it was discovered that he had been diagnosed as having pulmonary TB 7 months prior to this admission. The child had actually had a positive TB skin test and had been treated with isoniazid (INH) with 100 puis before the prescription ran out and was never refilled.
Dr. Katz: The average age of children who have TB of the spine is 9, so initially we were skeptical of this diagnosis. However, I have subsequently talked to physicians with a lot of third world TB experience who have seen numerous such cases.
Dr. Listernick: Why does TB have a predilection for the spine?
Dr. Katz: It's spread hematogenously, but I don't have a good answer.
Dr. Listernick: Can one have extrapulmonary TB without having evidence of lung disease?
Dr. Katz: Yes. TB most certainly was in the lung at some point, but you don't necessarily have to have clinical or radiologic evidence of pulmonary disease at the same time as extrapulmonary disease. Kids with tuberculous adenitis or meningitis don't always have abnormal chest x-rays. As many as 30% of children who have TB have extrapulmonary disease; that percentage is considerably lower in adults.
Dr. Chadwick: The vast majority of cases of TB vertebral osteomyelitis, or Pott's disease, have negative chest x-rays. The primary pulmonary process occurs so long before the development of osteomyelitis from hematogenous spread that cell-mediated immunity has contained the pulmonary infection successfully.
Dr. Curran: This child had a very small left lower lobe infiltrate, which may be an area of atelectasis. There was no hilar adenopathy or calcifications.
Dr. Listernick: This child had an open biopsy of the spine and soft tissue mass.
Pauline Chou, MD, pediatric pathologist: The bony fragments show areas of necrosis with inflammatory cells. In addition, there are areas of caseous necrosis with a few giant cells, but no well-formed granulomata. Within these necrotic areas, we can see several acid-fast bacilli, consistent with Mycobacterium tuberculosis. Of course, we'll need to wait for the culture to know for sure.
Dr. Listernick: So, it looks like she definitely has TB of the spine. What would you have done if the biopsy had not been as revealing?
Dr. Chadwick: We still would have treated her for TB. This is a classic story, notwithstanding the very young age.
Dr. Nordli: Is there a role for steroids in this child, knowing now that she has TB?
Dr. Chadwick: Steroids are important in highly inflammatory forms of TB, such as TB meningitis or endobronchial TB that is obstructing the airway. If the cord compression were due to edema or to a large mass of inflammatory tissue, steroids could have been helpful. However, if you administer steroids in situations where TB is suspected, it is crucial to start antituberculous therapy as well to prevent hematogenous spread. It's generally the rule of thumb that definitive antituberculous therapy is started long before TB has been proven by culture.
Dr. Listernick: How should she be treated?
Dr. Katz: First, there's no evidence that TB osteomyelitis requires a greater number of drugs than does pulmonary TB. The real issue is duration of therapy. For TB osteomyelitis, patients should probably receive at least a year of therapy with four antituberculous drugs. Twentyfive years ago, standard therapy for pulmonary TB was isoniazid and rifampin for 9 months; this produced a cure rate of greater than 95%. In the 1980s, we learned that 6 months of three-drug therapy was equivalent. Next, high resistance rates to both isoniazid and rifampin developed in certain areas, leading to the recommendation that four drugs be used for initial treatment if the sensitivity pattern is not known. Some infectious disease specialists start with three drugs if the incidence of TB in the community is low, suggesting that there is less of a likelihood of resistant strains.
Dr. Listernick: What drugs are used?
Dr. Katz: Isoniazid, rifampin, and pyrazinamide are good bactericidal drugs. Picking a fourth drug becomes somewhat tricky. One choice, streptomycin, has to be given intramuscularly and is ototoxic. In the doses that are used, ethambutol is bactericidal but may be toxic to the visual system. It's difficult to monitor the visual acuity of very young children who receive ethambutol.
Dr. Listernick: Why did this girl stop taking her medication?
Dr. Katz: We don't know. The father was very compliant with his therapy. The Department of Public Health will send out workers to administer therapy daily to patients with active TB (Directly Observed Therapy, or DOT). Unfortunately, they don't have the manpower to perform DOT on all positive skin test converters with latent or asymptomatic TB.
Dr. Listernick: How was she treated?
Dr. Chadwick: The previous monotherapy complicated this case, making the likelihood of an INH-resistant organism even higher. She was treated with four drugs: isoniazid, rifampin, pyrazinamide, and ethambutol. She'll receive complete ophthalmologic examinations every 3 months. The first sign of toxicity from the ethambutol is loss of color vision, which may be difficult to detect in such a young child. Eventually, we hope to culture the organism and identify its sensitivities so that we can narrow our therapy.
Dr. Listernick: The culture ultimately grew Mycobacterium tuberculosis. The sensitivities are pending. Thank you, everybody.