P&S Medical Review: Oct 1994, Vol.2, No.1
Clinicopathological Conference: A 41-year-old Man with Human Immunodeficiency Virus infection Presents with Seizures

GIORGIO CATTORETTI, M.D., JAY F. DOBKIN, M.D., PAMELA U. FREDA, M.D., JEFFREY A. MILLER, M.D., AND ABBIE I. KNOWLTON, M.D.
Columbia University College of Physicians and Surgeons, New York, NY

PRESENTATION OF CASE

This 41 year old Peruvian ex-taxi driver with known Human Immunodeficiency Virus infection presented to the Columbia Presbyterian Medical Center (CPMC) complaining of three seizures earlier in the day. In the Emergency Department, a fourth seizure was witnessed by the staff. He also complained of shortness of breath, chills, a productive cough, and one to two loose stools daily. His medications included hydrocortisone, flurocortisone acetate, trimethoprim-sulfamethoxazole (TMP-SMX), ciprofloxacin, fluconazole, and prochlorperazine. He was admitted for observation and treatment.

The patient first presented to CPMC thirty six months prior to admission with Pneumocystis carinii pneumonia. His total CD4 count at the time was 5 per cubic millimeter. He was discharged on a regimen of zidovudine and TMP-SMX. Over the next two years he was rehospitalized for diagnosis and treatment of Kaposi's sarcoma, cryptosporidiosis, and a gastrointestinal ulcer. During this time his sputum cultures grew Mycobacterium avium-intracellulare (MAI, or MAC for M. avium complex).

Six weeks prior to admission he was hospitalized with diarrhea, dehydration, and a temperature of 103o F. He had lost 14 kg over the preceding 16 months. His electrolytes were abnormal: sodium 109 mEq/L, potassium 4.0 mEq/L, blood urea nitrogen (BUN) 7 mg/L, and creatinine 0.8 mg/L. Workup established adrenal insufficiency: serum ACTH 379 pg/mL, repeat ACTH 453 pg/mL, serum aldosterone less than 2 pg/mL, and a fasting cortisol of 6.0mcg/dL that rose to 6.2 mcg/dL after cortrosyn stimulation. His electrolytes improved with rehydration and steroid replacement. He was discharged from the hospital on hydrocortisone.

Born in Peru, the patient immigrated to the United States several years ago. He is married and has two children. He denies alcohol abuse, intravenous drug abuse, and homosexual activity but reports frequent sexual intercourse with prostitutes. In addition to the above, his past medical history is significant for a seizure during his penultimate admission, and syphilis that was reportedly treated.

TABLE 1

---

Variable	On Admission
White-cell count (per mm3)	3,400
differential count (percent)
Neutrophils Band forms Lymphocytes Atypical lymphocytes Monocytes	75 1 17 4 3
Hemoglobin (mg/dL)	7
Hematocrit (percent)	21
Platelet count (mm3)	191,000

On physical examination, the patient was a cachectic man with temporal wasting. Vital signs in the Emergency Department included temperature 103.8 F, blood pressure 100/60 mmHg, pulse 104/min, and respirations 28/min. His weight was 50 kg. Fundiscopic exam revealed a whitish-yellow exudate on the right. The left fundus was not visualized. Examination of the chest, heart, and abdomen was not significant for any abnormal findings. Small inguinal nodes were palpable. Stool was Guiaic negative. The patient was oriented to person and place, and no focal neurologic abnormalities were identified. Laboratory values on admission are summarized in Tables 1-3.

Table 2

Variable	On Admission
Urine specific gravity	1.005
pH	6
Protein	0
Glucose	0
Acetone	0
Red-cell count (cells/hpf)	0-1

HOSPITAL COURSE: The patient was rehydrated intravenously with normal saline, started on hydrocortisone (100 mg every 8 hours), ampicillin, gentamicin, and transfused 2 units of packed red blood cells. His temperature ran a labile course, ranging from 98.6 F to 104 F daily. His sodium rose to 150 mEq/L and despite potassium supplementation, potassium fell to 3.0 mEq/L. Both returned to baseline levels following a steroid taper.

TABLE 3

Variable	On Admission
Sodium (mEq/L)	129
Potassium (mEq/L)	3.4
Chloride (mEq/L)	105
Carbon dioxide content (mEq/L)	11
Urea Nitrogen (mg/dL)	9
Creatinine (mg/dL)	1.2
Glucose (mg/dL)	166
Magnesium (mEq/L)	1.7
Calcium (mg/dL) Ionized calcium (mg/dL)	7 1.05
Protein (g/dL) Albumin (g/dL)	4.3 2.2
Total bilirubin (mg/dL)	0.2
Alkaline phosphatase (U/L)	169
Aspartate aminotransferase (U/L)	41
Alanine aminotransferase (U/L)	31
Lactate dehydrogenase (U/L)	246

On the 10th hospital day he became tachycardic and tachypneic. A chest roentgenogram demonstrated a nodule in the right lung. An arterial blood gas profile consisted of pH 7.48, pCO2 22 mm Hg, pO2 102 mmHg. A V/Q scan was indeterminate for a pulmonary embolus. On hospital day 15, he had shaking chills and abdominal discomfort. His white blood cell count was 3,600 with a left shift (86 percent polymorphonuclear leukocytes). Liver enzymes included alkaline phosphatase 445 U/L, aspartamine aminotransferase (AST) 246 U/L, alanine aminotransferase (ALT) 83 U/L, and an amylase of 68 U/mL. An abdominal ultrasound demonstrated an echogenic liver with cysts in the right lobe, no biliary obstruction, and normal kidneys. He was restarted on high dose hydrocortisone followed by a subsequent rise in sodium. AST and ALT levels returned to baseline. However, alkaline phosphatase remained elevated.

On the 23rd hospital day, Herpes simplex virus was cultured from a labial lesion and acyclovir was begun. On day 25, he suffered another seizure. A head computed tomography (CT) scan demonstrated a small lucency in the left caudate. A lumbar puncture yielded no red or white blood cells, protein 26 mg/dL, and glucose 30 mg/dL (serum glucose 264 mg/dL). The cerebrospinal fluid was culture negative, cryptococcal antigen negative, RPR negative, but positive for toxoplasma IgG and negative for toxoplasma IgM.

The patient developed patchy densities on chest roentgenogram. Alkaline phosphatase continued to rise to 1,127 U/L with no evidence of biliary dilatation by ultrasound. On his 48th hospital day, he suffered another seizure and was subsequently unresponsive. He died on his 49th hospital day.

DIFFERENTIAL DIAGNOSIS

DR. DOBKIN: Clearly with so much happening to this patient, the challenge is not only to try to reach a diagnosis, but to target the evaluations and treatment to things that are most valuable to the patient, in terms of morbidity and survival. It is possible within this category of patients to work them up almost endlessly. Therefore, a practical focus is quite important.

This patient presents with an enormous array of potential complications and disease manifestations. Table 4 is a representative list, taxonomically organized and by no means complete, of the possible agents that may be involved as opportunistic infections in AIDS. We have already mentioned in the case presentation evidence of past or current infection with cryptosporidium, Pneumocystis carinii, Mycobacterium avium intracellulare, as well as Kaposi's sarcoma. I would add histoplasmosis, Toxoplasmosis gondii, and infection with cytomegalovirus and Mycobacterium tuberculosis to this man's differential diagnosis list as well.

Table 4 Representative Opportunistic Infections in AIDS

With this enormous array of opportunistic infections, we certainly need some guidelines as to what to work up. There are some useful criteria available. Figure 1(not shown in on-line version) is a schematic of what we think is the natural history of HIV disease. HIV infection is a progressive, subacute to chronic process. On average, even with no medical intervention, over 10 years elapse between infection with HIV and the first appearance of an AIDS-defining illness such as those listed in Table 4. With the medications currently available, especially for prophylaxis of opportunistic infections, this time frame has been extended substantially.

But how do we know where an individual patient stands? The immunodeficiency at the root of AIDS was recognized very early on as being well-reflected by the number of circulating CD4-positive T lymphocytes. This test, called a AT4@ or ACD4 count@ remains the cornerstone of patient management. Thanks to data from a variety of sources, especially several excellent prospective studies of the natural history of HIV disease, we now know a good deal about the prognostic value of the CD4 count. As Figure 2 (not shown in on-line version) demonstrates, once the CD4 count drops below about 200 cells per cubic millimeter, the likelihood of opportunistic infections increases sharply.

PCP is the disease best defined by the CD4 count. About one-third of patients with T4 counts below 200 cells per cubic millimeter will develop PCP over 36 months if not given prophylactic treatment. Although the relationship of CD4 count to the other complications is not always as clear, some are fairly well defined. Disseminated infection with MAI tends to be associated with the lowest CD4 counts (less than 50) and ordinarily is not a presenting feature in AIDS patients since PCP usually occurs first. These relationships provide guidelines that can be helpful in the differential diagnosis of ill patients, as well as for staging and managing asymptomatic patients.

Table 5 reflects the clinical approach to primary care of HIV infected patients based on the CD4 count. At CD4 counts greater than 500 cells per cubic millimeter, there appears to be little or no indication for HIV-specific intervention. Screening for TB, immunizations, and addressing other health needs such as psychological and social issues are clearly important. As the CD4 count declines the need for other interventions increases. Anti-retroviral therapy has been controversial and complicated; there may or may not be a clear indication for initiation of anti-retroviral therapy in the CD4 range of 200-500 cells per cubic millimeter. A CD4 count of 200 per cubic millimeter is clearly a major milestone, and is very clearly established as the threshold of risk for symptomatic Pneumocystis infection. Pneumocystis prophylaxis has been demonstrated to be highly effective, contributing substantially to the decline of overall mortality of patients with AIDS in the last several years. Unfortunately, preventing earlier opportunistic infections does not correct the underlying immunodeficiency; others will be encountered later. The patient presented here well illustrates this phenomenon.

Table 5

Let us now look in detail at the current case. The patient is Peruvian; this can have several implications. Toxoplasmosis, a major cause of focal brain disease in AIDS, appears to result from reactivation of old infection. The risk of Toxoplasmosis is strikingly different around the world. There are much higher rates in poorer countries and in the poorer subpopulations, implying that environment is important for exposure to occur. Therefore, knowing the geographic background in a patient is important. Tuberculosis exposure is a function of environment and socioeconomic status. Histoplasmosis is also geographically restricted; it is very common in the Ohio River Valley of the United States, South America, and the Caribbean, but virtually never seen in lifelong New Yorkers who have not traveled to endemic areas. These three diagnoses must be considered higher on the list than some others.

Another consideration is the impact of prophylaxis on the risk of various opportunistic infections. TMP-SMX prophylaxis for PCP seems to decrease the risk of toxoplasmosis, while aerosolized pentamidine does not. Fluconazole given to treat esophageal candidiasis may decrease the chance of developing cryptococcal meningitis. However fluconazole resistant cryptococcal infection may then occur.

After his initial discharge on fluconazole and TMP-SMX, the patient returned with abdominal pain and was found to have gastric Kaposi's sarcoma and crytosporidiosis. We are unsure if cryptosporidia were found on endoscopy or from stool exam. If he had upper gastrointestinal cryptosporidiosis, he could have developed a syndrome of sclerosing cholangitis associated with cryptosporidia in the biliary tree. While I do not think he had that, it is worth considering in a patient who presents with abdominal pain and an obstructive pattern of liver function tests.

Sputum culture was positive for MAI, a common saprophyte in normal sputum specimens. However, in patients with advanced HIV disease, recovery of MAI from sputum may have prognostic significance as an early indicator of disseminated infection. One feature of disseminated MAI is massive hepatosplenomegaly. However, since this patient's abdominal symptoms improved with fluconazole it is unlikely that MAI was the underlying etiology.

Figure 3 MRI of head demonstrating dilated ventricles and prominent sulci. Note lucency in left internal capsule.

Two months later repeat endoscopy demonstrated that his Kaposi's sarcoma was gone, perhaps illustrating an interesting biological point: Kaposi's sarcoma is not a true malignancy. It behaves as a cytokine-regulated neoplasm. Products of the Tat gene in tissue culture can transform fibroblasts into KS cells. In transplant patients whose immunosuppression is discontinued, spontaneous regression of KS is well known. This is very exciting because it offers the potential for treatment by cytokine inhibition. There is nothing about this patient's course which indicates that his immunodeficiency improved, but it is conceivable that this accounts for the lack of Kaposi's sarcoma on repeat endoscopy.

The first indication that there might be endocrine dysfunction was the patient presentation with a sodium of 129 mEq/L. The differential diagnosis for this presentation of a patient with a history of diarrhea and this hyponatremia could include iatrogenic volume replacement with hypotonic fluids. However, in AIDS, there are other important issues. The syndrome of inappropriate anti-diuretic hormone secretion (SIADH) is very common in AIDS patients because of the frequency of pulmonary and intracranial lesions, however, we do not have evidence for either of these abnormalities.

That brings us to the final admission, in which the initial findings were basically non-specific. Given the seizure presentation, the findings on the magnetic resonance image (MRI) and CT scan are important. This presentation is typical for a patient like this. With advanced immunodeficiency, a history of immigration from a high-risk area, and a focal CNS presentation, a diagnosis of toxoplasmosis should be presumed and treated empirically. The history of TMP-SMX might decrease the likelihood of infection, but not eliminate the possibility. The positive CSF serology for toxoplasmosis certainly supports the diagnosis. The lumbar puncture ruled out cryptococcal meningitis, but not the rarer entity of CNS cryptococcoma. In terms of the seizures and the neuro-radiologic focal abnormalities, toxoplasmosis seems most likely. Based on the advanced stage of his HIV disease, other considerations would include CNS lymphoma, tuberculoma, or progressive multifocal leukoencephalopathy.

DR. KNOWLTON: Thank you Dr. Dobkin. Dr. Miller, would you comment on the radiologic aspects of the presentation.

DR. MILLER: The initial chest roentgenograms were from October 1991, the first admission when he came in with Pneumocystis carinii pneumonia. In the lungs, there are diffuse airspace opacities, a non-specific finding, created by the filling of the alveoli with fluid, cells, pus, or blood. This appearance is also classic for a fairly late stage of PCP. In comparison, early PCP often has a granular, ground glass appearance, especially around the hila. These kinds of airspace changes, however, strongly suggests a late stage of PCP.

A CT scan of the abdomen with contrast when he presented with Addison's syndrome demonstrated normal-appearing adrenal glands. Acutely, patients can present with TB of the adrenals; large adrenal glands that look puffed out, with non-concave borders. In later-stage Addison's, adrenal atrophy would be expected.

A MRI of the head (Figure 3), cut axially at the level of the lateral ventricles, demonstrates ventricles that are mildly dilated for a young man. This presentation is common in all AIDS patients because HIV, besides infecting the immune system, also infects the central nervous system, resulting in atrophy. The sulci are larger and more prominent than normal because of this atrophy. Also, there is a lucency of the left internal capsule that is not present on the right. This is one of the lucencies discussed in the case presentation which could act as a seizure focus. In the cerebellum (Figure 4) there is a small hyperintensity on the right, which could be a small infarct, an abscess from tuberculosis, or early lymphomatous lesion. Similar hyperintensities are visible in the right and left cerebellar hemispheres (Figure 5). While the cerebellar lesions are unlikely to cause seizures, they are abnormal and could be an infectious or neoplastic process.

Figure 4 MRI of cerebellum. Note hyperintensity in right cerebellum.

DR KNOWLTON: Thank you Dr. Miller. Dr. Freda has done some very interesting studies in the past year on patients with AIDS who have developed adrenal insufficiency. I though it would be of interest to hear the results.

Dr Freda: Many patients with AIDS have signs and symptoms which could be consistent with adrenal insufficiency, such as weakness, weight loss and hypotension. In addition hyponatremia occurs commonly in these patients and adrenal insufficiency should be considered in its differential. However, despite the frequent involvement of the adrenal glands of these patients with infectious and malignant processes, adrenal insufficiency has rarely been reported. Recently, we have documented primary adrenal insufficiency in five AIDS patients; this patient was one of these cases1.

Figure 5 MRI of cerebellum. Note hyperintensities in right and left cerebellar hemispheres (arrows).

This patient complained of weakness, nausea and diarrhea and had profound hyponatremia and hyperkalemia which led to the suspicion of adrenal insufficiency. The evaluation was then directed at distinguishing between primary and secondary adrenal insufficiency. This patient had an endocrinologic evaluation classic for primary adrenal insufficiency: an elevated plasma ACTH and low baseline cortisol with lack of rise after ACTH administration. The plasma ACTH level were very high at 379 pg/mL and 453 pg/mL, and baseline serum cortisol was 6.0 mcg/dL, which rose to only 6.2 mcg/dL after cortrosyn stimulation.

There are a number of possible etiologies for adrenal insufficiency in patients with AIDS. Primary adrenal insufficiency may be due to infiltration with infectious processes such as cytomegalovirus or mycobacterium as well as malignant processes such as lymphoma. These patients are also susceptible to adrenal hemorrhage or autoimmune adrenal disease. In addition drugs such as ketoconazole may lead to hypoadrenalism. Secondary adrenal insufficiency may also occur in AIDS patients and may be due to pituitary infiltration or pituitary suppression after glucocorticoid treatment.

AIDS patients with signs and symptoms consistent with adrenal insufficiency should be evaluated with a cortrosyn stimulation test and a baseline plasma ACTH concentration. Those patients with confirmed inadequate response should then be given glucocorticoid and mineralocorticoid replacement as needed.

The explanation for the recent increased incidence of adrenal insufficiency in AIDS patients is unclear. However, it may be that improved therapies for the infectious complications of this disease prolong the life span of these patients and the previously subclinical processes in the adrenal glands are given sufficient time to produce adrenal failure.

Figure 6 Low power view of adrenal gland demonstrating thickened, fibrotic capsule surrounding island of residual functioning glandular tissues (arrows)

DR. KNOWLTON: Thank you Dr. Freda. Dr. Cattoretti, would you comment on the pathologic findings.

DR. CATTORETTI: Pathology did not have permission for examination of the head, therefore some questions cannot be answered.

Both left and right adrenal glands were at the lower end of the average weight distribution for male adults: 3.6 and 4.2 grams respectively. The adrenal gland had little spared functioning parenchyma. On low power view, (Figure 6) we see that the adrenals are mostly replaced by fibrous tissue, with a markedly thickened, fibrotic capsule surrounding an island of residual cortex and medulla in the center (arrows). A high-power view of adrenal parenchyma (Figure 7), demonstrates CMV infected epithelial cells with both large nuclear inclusions (large arrows) and pink cytoplasmic inclusion bodies (small arrow). The infection produced loss of parenchyma with resulting adrenal insufficiency.

Figure 7 High power view of adrenal parenchyma with epithelial cells with large nuclear inclusions (large arrows) and cytoplasmic inclusions (small arrows).

Aspergillus was found invading the trachea and esophagus, with hyphae growing into the larger vessels, thrombosing them. A fungal abscess was seen in one section, which at high magnification (Figure 8) demonstrate hyphae growing outward, leaving a rim of necrosis and fibrosis. Aspergillus hyphae were destroying the bronchial wall and invading the wall of lung vessels, causing thrombosis and most likely septic emboli. This may have caused the some of the lesions on the MRI of the brain, which may have been aspergillomas. Diffuse CMV infection of the lung was also seen, typically accompanied by intraalveolar hemorrhage. No AFB positive specimens were seen.

Figure 8 Pulmonary fungal abcess at high magnification. Note hyphae growing outward and rim of necrosis and fibrosis (arrows).

The liver was grossly unremarkable. On microscopic examination a few areas of cholestasis were seen, although the bilirubin levels were within normal limits. In addition CMV infection of hepatocytes was seen.

Kaposi's sarcoma was present in the antrum of the stomach, with atrophic changes present in the surrounding mucosa. At higher magnification, the typical anastomosing channels made of newly formed vessels with atypical endothelial cells could be seen. Kaposi's sarcoma was noted in a biopsied para-aortic lymph node. Therefore, it had already extended outside the boundary of the stomach. In other parts of the same lymph node MAI was demonstrated by AFB stain.

Figure 9 Renal glomerulus in high power view. Note glomerular epithelial cells with nuclear inclusions (3 arrows).

CMV infection in the sub-mucosa of the entire GI tract was seen. The virus is present in the small bowel, cecum, and in the colon. In the colon, foci of aspergillus were also growing into the sub-mucosa.

The kidneys were also affected by CMV. In a high power view of a renal glomerulus (Figure 9), glomerular tuft cells with nuclear inclusions are seen (3 arrows). No inclusion bodies in the tubules were noted, although a patient can have CMV infection in the glomeruli without infection of the tubular epithelial cells.

In the spleen, many granulomatous lesions consisting of macrophages with fine, bluish, distended cytoplasm were seen. AFB staining demonstrated acid-fast bacilli filling the macrophages. These lesions were also seen in one of the two lymph nodes sampled.

DR. KNOWLTON: It is unfortunate that we do not have the pathology for the processes occurring within the head.

DR MILLER: One of the important points to make is the underlying importance of an autopsy in a patient whose death is not surprising, even if you have a good idea of the general medical problems. This case illustrates that patients with AIDS may have 5 or 6 different processes going on.

This patient does not resemble the CNS aspergillus cases I have seen in the past which typically involve cortical infarction from blood vessel involvement. Aspergillus is not among the common HIV-related complications. It took many years before it was recognized and seen at all. Many believe that the complication is promoted by steroid exposure. In many of the cases we have seen, steroid use for toxoplasmosis or syphilis was identified. However, in this case there was clearly no alternative because of the adrenal insufficiency.