P&S Medical Review: Nov 1993, Vol.1, No.1
A 53-Year-Old man admitted in alcohol withdrawal develops acute dyspnea on the tenth hospital day

Clinicopathological Conference

Steven Greenberg, M.D., Kevin A. Carnevale, M.D., Jay H. Lefkowitch, M.D., John H. M. Austin, M.D., and Abbie I. Knowlton, M.D.
Columbia University
College of Physicians and Surgeons, New York, NY

Presentation of Case

A 53-year-old Hispanic man with a six-year history of ethanol abuse presented to Columbia-Presbyterian Medical Center with tremors.

The patient was in his usual state of good health until one day prior to admission when he presented to the emergency department complaining of leg pains after hitting his legs on a chair. He was tremulous, with alcohol on his breath. He was afebrile with a pulse of 80 and a blood pressure of 110/90 mm Hg. He was alert and fully oriented without hallucinations or delusions. No abnormalities were noted on examination. The blood alcohol level was 389 mg/dL and the hematocrit was 43.7 percent. Five percent dextrose in normal saline was administered by vein. Injections of thiamine, folate, and chlordiazepoxide (25 mg) were given. Five hours after initial evaluation, the patient's condition had improved. He was discharged with a prescription for chlordiazepoxide (25 mg three times a day for three days), a referral to a detoxification center, and instructions not to drink.

On the following day, the patient presented complaining of headache, vomiting, and diaphoresis. He denied recent alcohol consumption. In the emergency department, he became increasingly agitated and was admitted.

Born in the Dominican Republic, the patient moved to New York City in his 30s. After serving seven years in prison, he was released at age 47 when he began consuming a pint of rum daily. He had a 30 pack-year history of cigarette smoking. He had no known drug allergies. The patient had an episode of pneumonia one year prior to admission. He had no history of pancreatitis, ulcer disease, cirrhosis, seizures, delirium tremens, psychotic symptoms, or hospitalizations related to alcohol consumption.

Physical examination revealed an agitated, tremulous man speaking incoherently. The temperature was 100.6°F, the pulse was 130, the respirations were 26, and the blood pressure was 120 by palpation. The sclerae were anicteric, conjunctivae were injected, and there was a crusty discharge along his eyelids. The lungs were clear and the heart examination was normal except for faint heart sounds and tachycardia. Gynecomastia was present. The abdomen was soft and the liver spanned 10 to 12 cm, extending 2 cm below the costal margin in the mid-clavicular line. The spleen was not palpable and there were no abnormal masses. Stool was heme negative. No spider angiomata or palmar erythema was noted. Distal pulses were intact and no peripheral edema or asterixis was present. He was awake but not oriented. Laboratory tests were performed (Tables 1-3).

     Table 1: Complete Blood Count
Variable                 On Admission
--------------------------------------------
White-cell count (per mm3)      9,200
Differential count (percent)  
   Neutrophils                     65
   Band forms                       1
   Lymphocytes                     32
   Monocytes                        2
Hemoglobin (mmol/L)                15.1
Hematocrit (percent)               45.6
Mean corpuscular volume (mm3)     109
Platelet count (per mm3)      139,000
--------------------------------------------

          Table 2: Urinalysis
Variable                 On Admission
--------------------------------------------
Specific gravity                1.03
pH                              5
protein                         1+
glucose                         0
acetone                         1+
White-cell count (cells/hpf)  3-4
--------------------------------------------

          Table 3: SMAC
Variable                 On Admission
--------------------------------------------
Sodium (mmol/L)               138
Potassium (mmol/L)              4.9
Chloride (mmol/L)             105
Carbon dioxide (mmol/L)        29
Urea nitrogen (mg/dL)          11
Glucose (mg/dL)               109
Creatinine (mg/dL)              1
Total cholesterol (mg/dL)     210
Protein (g/dL)                  6.9
   Albumin                      4.5
   Globulin                     2.4
Total bilirubin (mg/dL)         1.4
Alkaline phosphatase (U/L)     81
Aspartate aminotransferase     54
Alanine aminotransferase      274
 Magnesium (mmol/L)             1.3
--------------------------------------------

On the third hospital day, he developed redness around the intravenous catheter site located in the left antecubital fossa for which he received warm compresses. That night, the temperature rose to 104oF with a pulse 140. Respirations were 30, and the blood pressure was 160/110 mm Hg. The area around the catheter was erythematous and tender with a palpable cord. Blood cultures were drawn and oxacillin was begun. The following morning, the arm was improved.

On the fifth hospital day, a chest radiograph revealed a possible right middle lobe opacity. On hospital day six, Staphylococcus aureus, sensitive to oxacillin, was recovered from the blood cultures drawn on the third hospital day. A repeat radiograph of the chest on day seven demonstrated an increased opacity in the right middle lobe (Fig. 2). The patient defervesced and showed signs of clinical improvement on the eighth hospital day. Chlordiazepoxide was tapered and then discontinued on day nine.

On the morning of the tenth hospital day, the patient became somnolent and appeared acutely ill. The respirations were 30 and labored, the pulse was 100, and the blood pressure was 120/palpable. On examination, coarse expiratory gurgles were noted. Cardiac examination was notable only for tachycardia. The previous intravenous site remained mildly erythematous and swollen. Laboratory values were significant for a white-cell count of 13,800 and serum potassium of 2.5 mmol/L. The arterial blood gases (ABG) were pH 7.42, pCO₂ 38 mm Hg, pO₂ 64 mm Hg, and bicarbonate 24 mmol/L.

Five hours later an electrocardiogram revealed a marked bradycardia (that briefly responded to epinephrine) followed by ventricular fibrillation. After a prolonged resuscitation attempt, the patient became asystolic and was pronounced dead.

Differential Diagnosis

Dr. Knowlton: The patient we will discuss today, a very ill man who unfortunately cannot give an adequate history, presents with delirium tremens (`DTs'). Patients with DTs are nearly always uncooperative and difficult to examine, making clinical evaluation challenging. This patient was treated effectively with sedatives, but developed a serious infection from which he seemed to be recovering and then suddenly died. What conclusion did the students from Dr. Baer's group reach?

Medical Student: We believe the immediate cause of death was a pulmonary embolism which precipitated cardiac arrest. Evidence for this is the IV site was described as erythematous with a palpable cord.

Dr. Knowlton: Will Dr. Goland's group comment on the patient's diagnosis?

Medical Student: We thought his course was consistent with Staphylococcus aureus sepsis.

Dr. Knowlton: Yet he appeared to be getting better... What did Dr. Stern's group conclude?

Medical Student: We believe the patient suffered from gram- negative septicemia resulting from undiagnosed aspiration pneumonia. It may have been iatrogenic in nature secondary to the large amount of chlordiazepoxide originally used for sedation. The patient presented in a state of agitation and delirium; he was volume depleted which resulted in secondary hyperaldosteronism depleting his potassium. A state of hypomagnesemia also existed which may have led to hypocalcemia. Just prior to death, the patient may have been in shock leading to the release of epinephrine which also drove potassium and glucose into the cells contributing to the hypokalemia. These electrolyte abnormalities may have precipitated the bradyarrhythmia that occurred just prior to death.

Septic shock may have contributed to pulmonary edema by increasing the capillary permeability and decreasing myocardial contractility. The infiltrated intravenous site does not support the diagnosis of pulmonary embolism and there were no signs of deep vein thrombosis noted on examination. Aspiration is more likely in the clinical setting of a poorly alert patient who was noted to be vomiting upon presentation. A right middle lobe infiltrate is suggestive of pneumonia. The patient's gurgling sounds on auscultation of the chest are consistent with adult respiratory distress syndrome (ARDS) resulting from aspiration pneumonia, rather than pulmonary embolism. The final arrhythmia was a bradyarrythmia rather than a tachyarrhythmia associated with pulmonary embolism.

Dr. Knowlton: Dr. Greenberg, would you please give your interpretation.

Dr. Greenberg: I would like to summarize the most salient features of the case. The patient was a 53-year-old male alcoholic with a history of smoking who was admitted to the hospital one day following acute alcohol intoxication. On physical examination he appeared delirious, tremulous, and uncooperative. He was febrile, tachycardic, and tachypneic. Except for gynecomastia, he was remarkably free of signs of chronic liver disease. Pertinent laboratory values on admission included a surprisingly high hematocrit with macrocytosis, a normal white blood cell count, mild thrombocytopenia, and an increased SGPT. His chest X-ray was reportedly negative.

Since delirium tremens usually occurs 72 hours or more after abstinence from alcohol,[1] and the patient's delirium occurred much sooner, this suggests that he was suffering from the less ominous condition of major alcohol withdrawal syndrome. While the distinction between major alcohol withdrawal syndrome and delirium tremens is not precise, the mortality of delirium tremens is about 15 percent, whereas major alcohol withdrawal carries a much lower mortality rate.[2] Regardless of this distinction, both conditions are very serious disorders. The treatment that he was given, chlordiazepoxide, thiamine, and folate, was appropriate in either case. Although any benzodiazepine would suffice to treat alcohol withdrawal, lorazepam is preferred because it has a short half life and its dosing is easier to control.[3]

The most probable cause of delirium in a patient with alcoholism is either acute alcohol intoxication or major alcohol withdrawal and possibly delirium tremens. However, other common and treatable causes of delirium may coexist (e.g., hypoglycemia, hypernatremia, and Wernicke's encephalopathy). Both alcoholics and diabetics, as well as malnourished individuals, are predisposed to hypoglycemia,[4] which can lead to seizures and serious brain damage. Although this patient had very little evidence of chronic liver disease, liver failure is common among alcoholics and can lead to hypoglycemia by causing defective gluconeogenesis and inadequate hepatic uptake of insulin which leads to high peripheral insulin levels.[5] Hypernatremia resulting from dehydration can also lead to acute confusional states. An alcoholic patient is predisposed to Wernicke's encephalopathy due to thiamine deficiency as well. It is especially important to address these conditions because they are so readily treatable.

Other causes of delirium include major medical disorders such as sepsis, pneumonia, uremia, or liver failure and toxic effects of various drugs, especially those that have anticholinergic side effects.[6] The patient's laboratory results rule out the diagnosis of most of these electrolyte and metabolic disturbances. In particular, indices of his liver synthetic function (such as serum albumin, cholesterol concentrations, and prothrombin time) and his lack of significant stigmata of chronic liver disease on physical examination attest to an absence of significant hepatic disease. The baseline albumin and cholesterol concentrations are probably less than those reported on admission since he had evidence of volume depletion. The evidence for this was the urine specific gravity of 1.030 and a hematocrit of 45.6 percent which is uncharacteristically high for an alcoholic patient. These values, both of which were on the high end of normal, would likely be below the normal range if the patient were euvolemic. In summary, the patient is an alcoholic without significant liver disease admitted with a diagnosis of major alcohol withdrawal.

Alcoholism is a multisystem disease that affects every organ system of the body. Physicians caring for alcoholic patients should anticipate the major sequelae of this disease. Although this patient did not manifest gastrointestinal hemorrhage, many alcoholics do. The common causes of hemorrhage in the alcoholic patient are alcohol-induced gastritis, co-existent peptic ulcer disease, Mallory-Weiss tear, and esophageal varices if portal hypertension is present. Aspiration pneumonia is always a concern in patients with an altered mental status; however, this occurs during periods of acute intoxication or delirium. If a patient is neither intoxicated nor delirious and appears to be clinically improving in the absence of diseases of the esophagus or upper airway, there is no reason to suspect that there is a predisposition to aspiration. By the tenth hospital day he was not intoxicated with alcohol and his alcohol withdrawal syndrome had resolved so it is reasonable to assume that his mental status was relatively normal. It would be surprising if this man developed aspiration pneumonia beyond the third hospital day.

On the third hospital day the patient developed phlebitis of the upper extremity. This condition is fairly common in people who have intravenous catheters, especially for extended periods of time. Usually this does not give rise to major sequelae, but on occasion patients develop septic thrombophlebitis which could lead to endocarditis.[7] In this case the patient did have a significant fever (104°F) and a palpable cord in his upper extremity. As a physician caring for such a patient, what would you do in a situation like this?

Medical Student: Warm compresses are the recommended therapy for superficial thrombophlebitis.

Dr. Greenberg: Correct þ the accepted therapy for superficial thrombophlebitis is to administer warm compresses locally and possibly to prescribe aspirin. It is not recommended to administer intravenous heparin anticoagulation since the risk of embolic events as a result of phlebitis of superficial veins is negligible, whereas the risks associated with anticoagulation itself are higher. Although pulmonary embolism has been documented to follow upper extremity thrombosis, this is usually in the setting of an indwelling catheter of a central vein rather than a peripheral intravenous site.[8]

Dr. Greenberg: What is the accepted therapy for deep venous thrombosis (DVT)?

Medical Student: Systemic heparinization is an accepted therapy for DVT.

Dr. Greenberg: Yes. Regarding anticoagulation, what are the risks associated with anticoagulation?

Medical Student: It is contraindicated in alcoholics. Dr. Greenberg: Is it truly contraindicated in alcoholics?

Medical Student: Most people feel that alcoholics are prone to falling so anticoagulation should be avoided in order to prevent major head trauma and possible central nervous system hemorrhage.

Dr. Greenberg: I think you are quite right in recommending caution when considering anticoagulation in patients predisposed to head injury. But, anticoagulation is rarely contraindicated if someone has a life-threatening illness such as DVT or a pulmonary embolism.

Medical Student: Is anticoagulation contraindicated in someone who is actively bleeding?

Dr. Greenberg: It is important to distinguish between relative and absolute contraindications to anticoagulation. In someone who is actively hemorrhaging, heparin is contraindicated. On the other hand, if you have very good evidence that a patient has a thigh DVT or a pulmonary embolus and has coexistent mild gastrointestinal bleeding, then this is only a relative contraindication to anticoagulation. If someone has hemoptysis from a pulmonary infarction secondary to pulmonary embolism, it may be necessary to anticoagulate the patient anyway to avoid future and possibly lethal embolic events.

I do not know whether this man was given prophylactic subcutaneous heparin on admission. The housestaff may have decided to withhold such therapy when the risk of head trauma was greatest. However, the decision to withhold prophylactic anticoagulation in a bed-ridden patient should be reassessed after the patient's mental status clears. The fact that he has no significant evidence of chronic liver disease makes it highly unlikely that he has esophageal varices. The possibility that he is at risk for alcohol- induced gastritis is not a compelling argument against prophylactic anticoagulation.

On the fourth hospital day, he had a temperature of 104°F and blood cultures were drawn. Would you treat him with an antibiotic at this point?

Medical Student: Yes, I would recommend a broad spectrum antibiotic such as a cephalosporin.

Dr. Greenberg: He was given oxacillin, which is not a broad- spectrum antibiotic. What organisms commonly colonize peripheral IV sites? Staphylococcus epidermidis and S. aureus.[9] Yeast is probably the next most common, followed by enterococcus and various gram-negative organisms.[9,10] However the choice of which antibiotic(s) to use depends greatly on how the patient appears clinically. If a patient appears septic or is hypotensive and has obvious thrombophlebitis, it would be inappropriate to withhold broad-spectrum antibiotic coverage for the sake of elegance. It is not stated whether the patient appeared septic; however, his prompt response to oxacillin suggests that this antibiotic was an appropriate one.

At this point, it would be helpful to examine the radiographs.

Dr. Austin: The chest radiographic examination on hospital day two is normal (Fig. 1). On day five, the technique is slightly different with a greater degree of rotation. The right hemidiaphragm is slightly elevated and there is some blurry opacity just above it, medially. The possible locations of this process are either right middle lobe versus right lower lobe. One obvious suspicion is that the process represents pneumonia. The right heart border is intact, which suggests that it is in the lower lobe. Could this just be a pattern of atelectasis? It is not discoid atelectasis which has well-defined margins. So, I am suspicious of an infiltrative process in the lung, such as a pneumonia or even conceivably a pulmonary infarct. By day seven (Fig. 2) the opacity is still present, but by day ten a transverse band of opacity is apparent at the right base, which resembles `discoid' or `plate- like' atelectasis. During splinting, for example, this pattern becomes evident when a flat zone of lung parenchyma above and roughly parallel to the diaphragm becomes airless.

Dr. Greenberg: Dr. Austin, do you see a silhouette sign?

Dr. Austin: Yes, on day seven (Fig. 2) the right heart border is obscured, suggesting disease adjacent to the right heart in the middle lobe.

What diseases can produce this chest radiographic pattern? We all know that plain films of the chest are not at all sensitive for the diagnosis of pulmonary emboli.[11] What we look for when we suspect pulmonary emboli, which you see approximately 3 percent of the time, is dilated central pulmonary arteries which are manifestations of acute pulmonary hypertension. This is an important sign which signifies a potentially significant disruption of the pulmonary arterial tree. But we very rarely see that sign and I do not see it here.

Approximately 40 percent of patients with pulmonary emboli develop a pleural effusions which are almost always small.[12] Since about 50 to 60 percent of patients with pneumonia develop parapneumonic effusions,[13] this is a very non-specific sign. I do not see evidence of pleural effusions in any of these films.

Dr. Greenberg: Is there evidence of oligemia in the right upper lung?

Dr. Austin: Yes, it does look lucent. However, I think it is just a question of technique. That is another radiographic sign we look for. In theory, this is a result of decreased blood flow to that portion of lung which would otherwise be perfused by the occluded pulmonary artery. This is another famous sign that we practically never see.[14]

Dr. Greenberg: The overall conclusion upon reviewing the radiographs is that they are relatively unimpressive which is interesting for several reasons. When a patient suffers a respiratory deterioration and may be bacteremic, as this patient was, one explanation may be that he seeded his right heart valves with Staphylococcus and developed acute bacterial endocarditis which led to septic embolization to his lung. If this were the case, I would expect to see peripheral nodules on chest radiograph, perhaps cavitating, as a manifestation of septic emboli.[15]

Alternatively, given the findings of `gurgles' on physical examination of the chest, such a patient may develop cardiogenic pulmonary edema from mitral valve endocarditis or conceivably adult respiratory distress syndrome (ARDS) secondary to bacterial septicemia. However, the fact that the chest X-rays never displayed evidence of pulmonary nodules or diffuse infiltrates argues against these diagnoses. I think the strongest evidence against sepsis is that he defervesced while on oxacillin þ by day nine his temperature had fallen to 100.6°F. The time course of a disease tells a lot about what the disease might be.

One of the possibilities raised was that the patient was being inadequately treated for an infection with gram-negative organisms that colonized his intravenous catheter. Although colonization of intravenous catheters by gram-negative organisms is uncommon, one careful study documented colonization of catheter tips 12 percent of the time; gram-negative organisms accounted for 21 percent of total isolates; when present, colonization of catheters by these organisms was associated with bacteremia 45 percent of the time.[10] The study did not address the incidence of sepsis in those patients that were bacteremic. Another study documented catheter colonization in 5.4 percent (Vialon) catheter tips or 7.5 percent (Teflon) catheter tips, with no gram-negative organisms isolated, regardless of the catheter material.[9]

The fact that the patient improved on an antibiotic regimen that would have had no effect on gram-negative organisms strongly argues against this possibility. If he developed gram-negative septicemia, the source was unlikely to be his peripheral IV site. Death from septicemia results from either hypotension and ensuing cardiopulmonary arrest or from progressive hypoxemia resistant to mechanical ventilation (ARDS). However, when the patient was acutely ill, his systolic blood pressure was 120 mm Hg. In addition, the chest radiographs showed no evidence of pulmonary edema. Therefore, gram-negative septicemia is an unlikely explanation for his respiratory deterioration.

Consider for a moment the differential diagnosis of acute respiratory decompensation in hospitalized patients. What are the causes of tachypnea in hospitalized patients?

Medical Student: Pulmonary embolus can present as tachypnea and tachycardia.

Medical Student: Another possibility is septicemia.

Dr. Greenberg: Correct þ other causes of tachypnea in hospitalized patients include pneumonia, mucus plugging, pneumothorax, and cardiac tamponade. Pneumothorax and tamponade are uncommon in hospitalized patients. Basically, any lung pathology can produce tachypnea. Fluid overload is a very important cause of shortness of breath in hospitalized patients; a typical patient being someone who is uremic and who missed a dialysis session or a post-operative patient who was given many liters of fluid intraoperatively. These are common causes of fluid overload which give rise to sudden dyspnea and tachypnea. All causes of left heart failure, either related to volume status or secondary to acute ischemic events such as myocardial infarction or an incompetent mitral valve, may produce cardiogenic pulmonary edema. In this patient, most of these are easily ruled out.

On the tenth hospital day, the patient became acutely ill. He was tachypneic with a respiratory rate of 30. Of note was that his ABG revealed a pH 7.42, a pCO₂ 38 mm Hg and a pO₂ 64 mm Hg. Let's examine each value separately. A pH of 7.42 and a pCO₂ of 38 mm Hg are normal. A pO₂ of 64 mm Hg, however, is grossly abnormal. Arterial oxygen tension is age related, but a normal pO₂ is in the 90's mm Hg. So he is hypoxemic. Who would like to comment on his acid-base status?

Medical Student: At first glance it looks normal, but the low serum potassium suggests that there may have been multiple acid- base disturbances leading to what looks like a normal ABG.

Dr. Greenberg: Correct, his serum potassium concentration was 2.5 mmol/L, which is profoundly low. Common causes of hypokalemia include volume contraction and alkalosis. In fact, alkalosis and hypokalemia almost always coexist. Profound hypokalemia should raise the suspicion of significant metabolic alkalosis. Hypokalemia is also associated with hypomagnesemia,[16] which is very common in alcoholics. Although the causes of hypomagnesemia associated with alcoholism are probably multifactorial, a recent study underscored the importance of alcohol-induced renal tubular defects leading to wasting of several electrolytes including magnesium and potassium.[17] A patient who is not on diuretics would be unlikely to develop a potassium of 2.5 mmol/L without severe magnesium depletion.

Despite these laboratory findings which would be expected to lead to metabolic alkalosis, his bicarbonate concentration was normal. Another surprise is the lack of apparent respiratory alkalosis in the setting of marked tachypnea. The patient had a pO₂ of 64 mm Hg and is entitled to hyperventilate on that basis alone. So, I would expect him to have a respiratory alkalosis. Since he does not show evidence of this, it is possible that his somnolence, which he developed during this acute event, was evidence of a low cardiac output, blunting his central respiratory drive.

Another explanation as to why he did not display respiratory alkalosis is that he was suffering from respiratory muscle fatigue; this may have been exacerbated by hypomagnesemia, which has been linked to respiratory muscle weakness.[18] Severe hypokalemia, on rare occasions, can produce myopathy and respiratory failure,[19,20] but we are not told of the patient's muscle enzymes during the period of hypokalemia. Given his serum potassium concentration, I would have expected him to have a profound metabolic alkalosis prior to this acute event; the fact that he did not show evidence of this may have reflected coexistent metabolic acidosis secondary to lactic acidosis. I would like to know if he had an anion gap at the time. In any case, this degree of hypoxemia would warrant close observation if not transfer to an intensive care unit.

Before asystole occurred, the cardiogram showed marked bradycardia. Does he have an electrolyte disturbance that would predispose to an arrhythmia?

Medical Student: Hypokalemia.

Dr. Greenberg: Absolutely, hypokalemia predisposes to ventricular tachyarrhythmias. However, he had bradycardia. We can not make too much of that since we do not know what his rhythm was before he became bradycardic. It is possible that the patient was in shock and developed ventricular tachycardia or perhaps suffered a myocardial infarction. At this stage of his illness, it is hard to draw firm conclusions regarding the bradycardia.

Overall, the two most likely diagnoses that led to the patient's death are pulmonary embolism and bacterial septicemia. Of the two, pulmonary embolism is far more likely since the patient was already improving on oxacillin and his chest radiograph was surprisingly benign.

Pulmonary embolism is seldom lethal unless it is massive.[21] Patients succumb to pulmonary embolism because of profound right heart failure.[21] This patient probably developed shock due to low cardiac output secondary to overwhelming right heart failure. Despite his upper extremity thrombophlebitis, the source of his pulmonary embolism was probably the deep veins of the lower extremities since these account for at least 85 to 90 percent of all cases of fatal pulmonary embolism as determined by autopsy studies.[22] Pulmonary embolism from upper extremity venous thrombosis has been reported; however, it usually occurs from central rather than peripheral venous catheters.[8]

In conclusion, the diagnoses include massive pulmonary embolism with right heart failure and cardiopulmonary arrest, deep venous thrombosis of lower extremities, and alcoholic fatty liver.

Dr. Knowlton: Dr. Carnevale, would you report the autopsy findings?

Drs. Carnevale and Lefkowitch: The main pulmonary artery was opened and showed a thrombus extending into the branching vessels of the right lung (Fig. 3). The right lower lobe is deeply congested, consistent with the early congestive phase of a pulmonary infarct. Histologic sections of the thrombus show red blood cells mixed with fibrin and platelets (Figs. 4 and 5). These alternating layers are classically known as `lines of Zahn.' The layering effect is one that occurs only in vivo, during life. Hence, the pathologist's observation of `lines of Zahn' in a thrombus is important diagnostic evidence of antemortem thrombus formation. Such a thrombus could originate in deep leg veins, pelvic veins, or other sites and embolize to the lungs. Once in the pulmonary artery, additional surface thrombosis on the embolus may lead to extension of the thrombus.

The thrombus was only partly adherent to the wall of the right pulmonary artery without organization (Fig. 4). Such thrombi are usually no older than several days. This is also corroborated by the sections of lung parenchyma (Fig. 6) which show only marked vascular engorgement, but no outright tissue infarction. Based on these gross and microscopic lung and pulmonary artery findings, it appears that there was a large pulmonary thromboembolus fairly close to the time of the patient's death, here associated predominantly with vascular engorgement of the right pulmonary vessels. For an infarct to have developed, the patient would have had to survive for at least several days after the pulmonary embolus. Other findings at autopsy included a gastric ulcer and fatty liver consistent with alcohol abuse.

Dr. Greenberg: Did you examine the deep veins of the extremities?

Drs. Carnevale: Unfortunately, consent was obtained for only a limited autopsy, with studies limited to the abdomen, thorax and head. However, the pelvic veins were free of thrombi.

Dr. Greenberg: Was the vena cava patent?

Dr. Carnevale: Yes.

Dr. Knowlton: Did the arm appear well-healed?

Dr. Carnevale: Yes it did. However, cultures of the wound did grow Staphylococcus aureus.

Dr. Greenberg: What did examination of the heart reveal?

Dr. Carnevale: The heart was slightly dilated, but there was no evidence of infarction or thrombus.

Dr. Greenberg: There are several lessons to be learned from this case. One is that the course of a patient's illness will often provide valuable clues to the diagnosis. The fact that the patient was improving from his transient bacteremia suggests that sepsis was not the cause of death. This increases the likelihood of another diagnosis.

Another important lesson is that pulmonary embolism is common, difficult to diagnose clinically, and can be fatal. It is worth noting that the sensitivity of any given symptom or sign in diagnosing pulmonary embolism is disappointingly low. For example, while subjective dyspnea and tachypnea are present in the majority of cases of pulmonary embolism, they are absent in as many as 15 percent of patients with documented pulmonary embolism.[23] The absence of phlebitis, mentioned today as evidence against pulmonary embolism, cannot be used to rule out a diagnosis of pulmonary embolism; it was present in only 41 percent of documented cases of pulmonary embolism in one large series.[23] Pulmonary embolism is particularly difficult to diagnose in the face of an abnormal chest radiograph; this occurs in over 70 percent of such cases.[24] The most common finding is a non-specific infiltrate.[24] Other findings, mentioned by Dr. Austin, include small pleural effusions, atelectasis, large central pulmonary arteries, oligemia, and pleural-based pulmonary infarcts.[25]

Why is it important to note how common abnormal chest radiographic findings are in pulmonary embolism? A normal or minimally abnormal chest radiograph in a patient whose respirations are compromised should raise the suspicion of a pulmonary embolism. But an abnormal chest radiograph, for example one with an infiltrate, should not dissuade one from pursuing this diagnosis since this is a common finding in this disease. Pulmonary embolism is most often overlooked in the presence of concurrent lung disease.[26]

Anatomical Diagnosis

Massive acute pulmonary embolism.
Pulmonary infarct.
Steatosis of the liver.
Gastric ulcer.

References

McMicken DB. Alcohol withdrawal syndromes. Emerg Clin N Am 1990;8:805.

Sellers EM, Kalant H. Alcohol intoxication and withdrawal. N Eng J Med 1976;294:757.

Hollister LE, Greenblat DJ, Rickels K, et al. Benzodiazepines 1980: current update. Psychosomatics 1980;21:1S.

Freinkle N, Singer DL, Arky RA, et al. Alcohol hypoglycemia: I. Carbohydrate metabolism of patients with clinical alcohol hypoglycemia and the experimental reproduction of the syndrome with pure ethanol. J Clin Invest 1963;42:1112.

Lee WM. Acute liver failure. N Eng J Med 1993;329:1862.

O'Brien JG. Evaluation of acute confusion (delirium). Primary Care 1989;16:349.

Maki DG. Infections due to infusion therapy. In: Bennet JV, Brachman PS, eds. Hospital infections. 3rd ed. Boston: Little, Brown, 1992:849.

Monreal M, Lafoz E, Ruiz J, Valls R, and Alastrue A. Upper- extremity deep venous thrombosis and pulmonary embolism. Chest 1991;99:280.

Maki DG, Ringer M. Risk factors for infusion-related phlebitis with small peripheral venous catheters. Ann Int Med 1991;114:845.

Cooper GL, Hopkins CC. Rapid diagnosis of intravascular catheter- associated infection by direct Gram staining of catheter segments. N Eng J Med 1985;312:1142.

Buckner CB, Walker CW, Purnell GL. Pulmonary embolism: chest radiographic abnormalities. J Thorac Imag 1989;4:23.

Taryle DA, Potts DE, Sahn AS. The incidence and clinical correlates of parapneumonic effusions in pneumococcal pneumonia. Chest 1978;74:170.

Bynum LJ, Wilson JE. Radiographic features of pleural effusions in pulmonary embolism. Am Rev Resp Dis 1978;117:829.

Westermark, N. On roentgen diagnosis of lung embolism. Acta Radiologica 1938;19:357.

Kuhlman JE, Fishman EK, Teigen C. Pulmonary septic emboli: diagnosis with CT. Radiology 1990;174:211.

Millane TA, Ward DE, Camm AJ. Is hypomagnesemia arrhythmogenic? Clin Cardiol 1992;15:103.

DeMarchi S, Cecchin EA, Basile A, Bertotti A, Nardini R, Bartoli E. Renal tubular dysfunction in chronic alcohol abuse þ effects of abstinence. N Eng J Med 1993;329:1927.

Dhingra S, Solven F, Wilson A, McCarthy DS. Hypomagnesemia and respiratory muscle power. Am Rev Resp Dis 1984;129:497.

Hermann RA, Mead AW, Spritz N, et al. Hypopotassemia with respiratory paralysis. Case due to renal tubular acidosis. Arch Int Med 1961;108:925.

Dorin RI, Crapo LM. Hypokalemic respiratory arrest in diabetic ketoacidosis. JAMA 1987;257:1517.

Gorham LW. A study of pulmonary embolism Part II The mechanism of death; based on a clinicopathological investigation of 100 cases of massive and 285 cases of minor embolism of the pulmonary artery. Circulation 1961;108:79.

Havig O. Deep vein thrombosis and pulmonary embolism. Acta Chirurgica Scand Suppl 1977;478:1.

Stein PD, Willis PW, DeMets DL. History and physical examination in acute pulmonary embolism in patients without preexisting cardiac or pulmonary disease. Am J Cardiol 1981;47:218.

Szucs MM, Brooks HL, Grossman W, Banas JS, et al. Diagnostic sensitivity of laboratory findings in acute pulmonary embolism. Ann Int Med 1971;74:161.

Kerr IH, Simon G, Sutton MB. The value of the plain radiograph in acute massive pulmonary embolism. Brit J Radiol 1971;44:751.

Goldhaber SZ, Hennekens CH, Evans DA, Newton EC, Godleski JJ. Factors associated with correct antemortem diagnosis of major pulmonary embolism. Am J Med 1982;73:822.