Correspondence
Clinical Problem-Solving: Identifying Addison's Disease
N Engl J Med 1996; 334:1403-1405May 23, 1996
- Article
To the Editor:
I stopped reading the Clinical Problem-Solving article by Keljo and Squires (Jan. 4 issue)1 when I reached the second line of the second column, which gives the erythrocyte sedimentation rate as 38 mm per hour. I was dissatisfied with the reasoning of the discussant in the first paragraph. None of the proposed differential diagnoses seemed appropriate. I started again and constructed my own differential diagnosis. My summary of diagnostic signs and symptoms included shortness of breath during exercise without wheezing or cough, loss of weight, amenorrhea, hypotension with an increased pulse rate, and a previous vaccination. I came to just one possibility — Addison's disease. My request for additional information at this stage would have concerned loss of appetite, nausea, the effect of exposure to sun, pigmentation of skin folds, orthostatic hypotension, the exact eosinophil count in circulating blood, and levels of glucose, sodium, and potassium.
My criticism of the discussant's differential diagnosis is that none of the diagnostic items mentioned are suggested by the complete set of findings known at that stage, except for autoimmune disease, if the discussant had the adrenals in mind. A body-mass index of 16.6 in a patient with anorexia nervosa would not give rise to amenorrhea and hypotension; girls with anorexia also tend to be overactive and not to complain about exertion. Hyperthyroidism is not characterized by hypotension and a moderate pulse pressure; neither are inflammatory diseases of any cause, unless both adrenals are involved. Malabsorption leading to the symptoms described would be evident from the medical history, and hypotension due to a thiamine deficiency in malnutrition should be accompanied by pulmonary edema. Malignant or inflammatory disease leading to the level of wasting described is almost certainly accompanied by many other important findings. Anemia in itself is not always associated with low blood pressure.
What went wrong in this clinical case? First, the patient's circulatory condition was overlooked. This aspect, combined with the absence of signs of severe general illness, leads to the diagnosis of hypocortisolism. Second, the discussant and the original clinicians constructed their differential diagnosis by using different parts of the available information for each item, neglecting signals that were contradictory or absent. Third, when there was no solution, they should have started again at the very beginning of the diagnostic process and followed the basic systematics of data collection and evaluation.
1 ReferencesWolter S. de Loos, M.D., Ph.D.
Central Military and University Hospital, 3509 AA Utrecht, the NetherlandsTo the Editor:
Dr. Keljo and Dr. Squires brilliantly discuss the case of a 15-year-old white girl with a delayed diagnosis of Addison's disease. Primary physicians, many consultants, and even an endocrinologist did not have a clue to the diagnosis, nor did they include Addison's disease in their long list of differential diagnoses. Many tests, even invasive ones, were performed in a very sick patient who was almost in adrenal crisis. The reason for the delay in diagnosis was the absence of increased pigmentation in the skin and mucous membranes. The question, however, is whether gray pigmentation in the mouth and buccal mucosa1 was missed or whether increases in brown pigmentation on the face and especially in pressure points, such as the elbows, knees, palmar creases, nipples, and areolae, were misinterpreted as an excessive suntan.2 In 92 percent of cases of Addison's disease, hyperpigmentation is found.2 It can even precede other manifestations of the disease, such as fatigue, weakness, progressive asthenia, and orthostatic hypotension.
The absence of increased pigmentation is very unusual in a patient with primary adrenal insufficiency, except in adrenal insufficiency due to pituitary failure, which was not the case here (this patient had a high level of corticotropin). This case teaches us a lesson: if a crucial sign is missing in a rare disease, the correct diagnosis can be very difficult and frustrating to reach and can be much delayed. But the persistence of the physicians who were caring for the patient paid off. Finally, just in time, the diagnosis was clinched, and the patient's life saved. A rare diagnosis that is obvious in retrospect is often not so obvious prospectively, however.
2 ReferencesDushan J. Babich, M.D.
P.O. Box 2577, Southampton, NY 11969To the Editor:
Regarding the clinical case “Just in Time,” I missed the diagnosis too. However, on rereading, I see that no differential count was reported (and perhaps not done?) at the time of any blood count. Had such a count shown increased levels of eosinophils and lymphocytes, said to be common in adrenal insufficiency, would some clinicians have gotten a clue many weeks and thousands of dollars earlier?
I have seen reports that “routine” differential counts are not “cost effective.” Well, maybe.
David G. Baxter, M.D.
351 E. Temple St., Los Angeles, CA 90012To the Editor:
In the case of a 15-year-old girl who reported fatigue and shortness of breath on exertion, after a month of evaluation and therapeutic trials aimed at various causes of fatigue, the invited “expert clinician” arrived at the diagnosis, Addison's disease. The authors conclude, “Only the toughest critic could fault any of the clinicians for not making the correct diagnosis earlier.” As with so many arguments, that depends on one's point of view. Or, as my grandfather put it, “Teach a man to use a hammer, and he will treat every problem as if it were a nail.” Both the expert and the authors appear to be gastroenterologists. An alternative analysis might have focused the evaluation on the symptom — shortness of breath on exertion — that the expert acknowledged is “an uncommon symptom in childhood.”
Shortness of breath indicates that unexpected work is required to move the flow of air or blood in order to meet the increasing demands for oxygen consumption and carbon dioxide removal with exercise. Thus, one would like to have measured ventilation and gas exchange during graded exercise in order to determine which physiologic variable was perturbed in this girl. Had this been done, the clinician would probably have discovered a cardiovascular limitation to gas exchange, with little change in the stroke volume and at best a doubling of the heart rate; the initial heart rate was 94 beats per minute. A two-dimensional echocardiogram would probably have revealed small ventricles and no evidence of cardiovascular disease. These findings would have raised questions about circulatory volume, and the appropriate diagnosis would probably have been “nailed” more expeditiously. Thus, the authors' conclusion that only the toughest critic could fault the clinicians depends on one's point of view and on whether one was taught to use a hammer or a saw.
Daniel C. Shannon, M.D.
Massachusetts General Hospital, Boston, MA 02114To the Editor:
The article by keljo and Squires presented an intriguing opportunity to test DXplain, a decision-support system designed to provide appropriate clinical information to physicians faced with puzzling cases.
I entered the data in the order they were presented. After the data obtained during the initial history taking and physical examination were entered, the list of diagnoses proposed by the system focused, as did the physicians, on (relatively) common disorders, such as anorexia nervosa and Graves' disease. When the epigastric pain and nausea and all the normal initial basic laboratory-test results were entered, chronic pancreatitis and several lymphomas moved to the top of the list, with chronic adrenal insufficiency now appearing as number 17. When I added the vomiting, the dehydration, and the second round of laboratory tests, Addison's disease moved up to number 2 — well before this disease was considered by the physicians caring for the patient or by the discussant.
This test shows the strength of a decision-support system: it suggests possible diseases to the physician, who is in a much better position to make a diagnosis once the disease has been considered.
Edward P. Hoffer, M.D.
Massachusetts General Hospital, Boston, MA 02114To the Editor:
In the Clinical Problem-Solving case by Keljo and Squires, little was said about the possible causes of Addison's disease, and nothing about the causative roles of histoplasmosis and tuberculosis. In these days of our awareness of problems with tuberculosis control, this subject seems worth a word or two. Thus, information about the area where this 15-year-old lives or lived, her living conditions, and the possibility of contact with tuberculosis becomes important.
Gerald L. Baum, M.D.
Israel Lung Association, Tel-Aviv 63346, IsraelAuthor/Editor Response
The authors reply:
To the Editor: The patient's physicians believed that the diagnoses considered did not explain all the features of her illness, and they continued to explore new avenues. Had the diagnosis of Addison's disease been entertained before her electrolyte levels became abnormal, it might well have been made earlier. The question at issue is, What pointers may have been missed?
We agree with Dr. de Loos and Dr. Shannon that the patient had evidence of hypovolemia throughout her illness, suggesting the diagnosis of Addison's disease. Earlier recognition of the chronic nature of the hypovolemia might well have speeded the diagnosis. Her cardiovascular status at the time of presentation did not, however, exclude malnutrition due to inflammatory bowel disease.
Amenorrhea is uncommon in Addison's disease, and if present it generally indicates cachexia.1 The presence of amenorrhea cannot therefore be used to distinguish Addison's disease from any other source of malnutrition. (Ovarian failure is also seen in patients with Addison's disease, but that clearly did not apply to this patient, whose menses returned after treatment.)
The immunization and the Addison's disease were not linked, because the amenorrhea began six months before the immunization was given. The family simply did not notice her illness until after the immunization.
Dr. Baum reminds us that tuberculosis was the most common cause of Addison's disease not so very long ago and that tuberculosis is becoming more common again. This patient had a normal chest film and a negative response to a purified-protein-derivative tuberculin skin test.
Multiple caretakers commented on the patient's large number of deeply pigmented nevi, and there is a report of such changes in Addison's disease.2 This child did not have the more classic dermatologic changes of Addison's disease brought up by Dr. Babich.
Dr. Baxter's question is appropriate. However, the patient's differential count was normal: 50 percent segmented neutrophils, 1 percent band forms, 2 percent monocytes, and 47 percent lymphocytes.
Dr. Hoffer's report on the performance of DXplain is impressive. As Dr. Shannon implies, clinicians work from short lists, and these short lists vary widely from specialty to specialty. When none of the diagnoses entertained explain the findings, the list must be expanded. DXplain provides one way to do this.
2 ReferencesDavid J. Keljo, M.D., Ph.D.
Robert H. Squires, Jr., M.D.
University of Texas Southwestern Medical Center at Dallas, Dallas, TX 752351
Bethune JE. The diagnosis and treatment of adrenal insufficiency. In: DeGroot LJ, Besser GM, Cahill GF Jr, et al., eds. Endocrinology. 2nd ed. Vol. 2. Philadelphia: W.B. Saunders, 1989:1647-59.
2
Ibsen HH ,Clemmensen O . Eruptive nevi in Addison's disease. Arch Dermatol 1990;126:1239-1240
CrossRef | Web of Science | Medline
