CPC #6: Answer

The most striking finding at autopsy was that of multiple small (one to two millimeter) pulmonary nodules scattered throughout the lungs bilaterally (Figure 1).  The nodules were slightly larger in the apical portions of the lung.  Microscopically, these nodules demonstrated vague granulomatous inflammation (Figure 2), but special stains for acid-fast bacilli were negative in these sections.  The hilar lymph nodes were matted together grossly, and microscopically demonstrated active caseating granulomatous inflammation.  On special stains, rare acid-fast bacilli were identified here, and more acid-fast bacilli were identified in granulomas involving the liver (Figure 3).  More relevant to the patient’s symptoms, examination of the brain revealed a cloudy exudate involving the base of the pons (Figure 4).  Microscopically, this corresponded to active caseating granulomatous inflammation and exudate involving the subarachnoid and subdural spaces (Figure 5 and 6).  Numerous acid-fast bacilli were identified on Kinyoun’s acid-fast stains in the subarachnoid space (Figure 7).  Additional findings in the brain included large areas of cortical necrosis secondary to herniation, and focal areas of necrosis in the pons associated with hemorrhage.  There was also evidence of renal glomerular disease (crescents) similar to that seen in the ante mortem renal biopsies, consistent with systemic lupus erythematosis. Other findings of note at autopsy included an organizing thromboembolism to the right middle lobe of the lung, and hyaline membrane formation as a near terminal complication.

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

The autopsy findings are those of miliary tuberculosis, which refers to hematogenous dissemination of tuberculosis from a primary focus.  This should be distinguished from cavitary fibrocaseous tuberculosis in the lung and mediastinal lymph nodes. Characteristically, patients develop multiple small pulmonary nodules, which are larger at the apex, possibly due to increased oxygen saturation in this part of the lung.    It is most likely that the patient had a prior tuberculosis infection, which was reactivated from the hilar lymph nodes due to immunosuppression, and disseminated hematogenously.  The fact that more mycobacteria were identified in the central nervous system than in the lung or hilar nodes likely reflects the fact that the CNS is essentially a sanctuary for infectious agents, isolated from much of the body’s humoral and cellular immune systems.

While the presentation was unusual, this patient did demonstrate many of the classic clinicopathologic findings of central nervous system tuberculosis.  While CNS tuberculosis may present as an abscess, the more common finding is that of a basilar meningitis and exudate.  The basilar location of the exudate leads to several clinical seqella.  First, due to obstruction of the flow of cerebrospinal fluid, patients with CNS tuberculosis may develop hydrocephalus.  Second, because the exudate entraps and obstructs penetrating arterioles at the base of the brain, patients may develop basilar infarctions. Third, because the cranial nerves exit from the base of the brain, patients may demonstrate loss of specific cranial nerve functions as the disease progresses.  Our patient demonstrated all of these findings clinically and pathologically. 

The differential diagnosis for this case can essentially be divided into that of the initial presentation with new onset seizures, and that of the later progression from confusion to obtundation to coma. 

First, in a patient with lupus with new onset seizures, the differential diagnosis includes neuropsychiatric lupus, thrombosis from antiphospholipid syndrome, and opportunistic infection.  Neuropsychiatric lupus is the best explanation for this patient’s initial presentation with seizures.  This syndrome clinically presents with encephalopathy, confusion and seizures, and pathologically is usually associated with mild perivascular chronic inflammation and microhemorrhages in the brain.  Neuropsychiatric lupus is best considered a microvasculopathy.  True vasculitis in the brain is rare, and patient’s symptoms may be out of proportion to the pathologic findings.   Antiphospholipid antibodies are present in 50 percent of patients with SLE, and predispose to both arterial and venous thrombosis.  This possibility is far less likely in this case, considering that the patient does not demonstrate an elevated APTT.  Finally, opportunistic infection could explain the initial presentation, but this seems very unlikely given the fact that the cerebrospinal fluid did not show an exudative pattern at this time. 

As the patient progressed to coma, the cerebrospinal fluid progressed to a more infectious picture with decreased glucose, increased protein, increased leukocytosis, and increased opening pressure.  In a patient who is immunosuppressed, infection is clearly the number one cause for this picture. Aside from the ever-present tuberculosis, toxoplasmosis would be a leading possible diagnosis in this patient who has lived in Nigeria.  Toxoplasmosis usually presents as a brain abscess, but can rarely present with a meningoencephalitis as we saw in this case.  Serology would be key in establishing this diagnosis.  Other opportunistic infections, including histoplasmosis and cryptococcal disease were also considered, as was the possibility of a catastrophic presentation of the antiphospholipid syndrome.  

In summary, this patient’s course demonstrates the ever-present nature of tuberculosis in medicine.  It demonstrates an unusual and subtle presentation of a relatively common disease, one that manifests itself with greater frequency in patients who are immunosuppressed. 

Return to Top