Scleroderma, Fasciitis, and Eosinophilia Associated with the Ingestion of Tryptophan

Patient 1 was a 67-year-old man who was well until July 1988, when pruritus developed on his back, neck, and arms, and he had dyspnea on exertion. He had a leukocyte count of 9.9×10⁹ cells per liter, with 30 percent eosinophils, and a Westergren erythrocyte sedimentation rate of 15 mm per hour. A chest film showed a leftsided pleural effusion and atelectasis at the left base. The leukocyte count in pleural fluid obtained by thoracentesis was 0.36×10⁹ cells per liter, with 40 percent lymphocytes and 60 percent eosinophils. Cytologic studies and cultures were negative. A transbronchial biopsy of the left lower lobe revealed an eosinophilic infiltrate. The patient was given prednisone (40 mg daily), and his dyspnea and pleural effusion disappeared.

Six months later, one week after the prednisone was discontinued, skin tightness was noted, and the patient was referred to the Medical University of South Carolina. He had had a myocardial infarction in 1976 and had smoked cigarettes in the past. He reported no history of Raynaud's phenomenon, fingertip ulceration, dysphagia, recent dyspnea, or weakness. Physical examination revealed taut skin over the chest, abdomen, back, arms, thighs, and legs, but no sclerodactyly; the remainder of the examination was normal. The leukocyte count was 7.2×10⁹ cells per liter, with 7 percent eosinophils, and the sedimentation rate was 42 mm per hour. An antinuclear-antibody test was negative. Microscopical examination of the nail-fold capillaries was normal — i.e., there was neither enlargement nor loss of capillaries.⁵ A chest film and electrocardiogram were normal, and pulmonary-function tests indicated moderate obstructive lung disease, with a diffusing capacity of 63 percent of the predicted value. The results of skin biopsy were consistent with scleroderma, and prednisone therapy was reinstituted (30 mg daily).

We then learned that the patient had begun taking tryptophan (2.0 g daily) in June 1988, one month before the start of his illness. It was discontinued in April 1989. His skin tightness worsened initially, but subsequently improved, and the prednisone dose was reduced to 20 mg daily. In July 1989, he underwent studies of tryptophan metabolism. In January 1990, two weeks after an exacerbation of his skin lesions, he died suddenly of an apparent myocardial infarction.

Patient 2 was a 44-year-old woman with a 12-year history of Addison's disease, for which she took replacement doses of hydrocortisone and fludrocortisone. In April 1988, tender, inflammatory lesions developed on her legs. She had been taking tryptophan (1.0 g daily) for eight months. In December 1988, she was seen by a dermatologist, who noted erythematous plaques on her trunk and lower extremities. Multiple skin biopsies showed mild, chronic, superficial and deep perivascular infiltrates suggestive of an urticarial eruption.

In February 1989, she was noted to have waxy plaques with peau d'orange changes of the skin of her legs. She had a negative test for antinuclear antibodies, a sedimentation rate of 25 mm per hour, and eosinophilia (10 percent eosinophils). Her chest film and electrocardiogram were normal.

She stopped taking tryptophan early in 1989. She was first evaluated by us in September 1989. She reported no history of Raynaud's phenomenon, fingertip ulceration, dysphagia, dyspnea, or muscle weakness, but she did report severe pain in her legs and arms. Physical examination revealed only ill-defined areas of hyperpigmentation and tender induration and swelling of the skin of the legs, thighs, arms, and trunk, but no sclerodactyly or telangiectasia. The leukocyte count was 9.3×10⁹ cells per liter, with 46 percent neutrophils, 2 percent monocytes, 13 percent lymphocytes, 38 percent eosinophils, and 1 percent basophils. Microscopical analysis of nail-fold capillaries was normal, as were a barium esophagogram and pulmonary-function tests. A skin biopsy showed fibrosis and inflammation with eosinophils and plasma cells.

In October 1989, the patient underwent studies of tryptophan metabolism. Subsequently, she was treated with prednisone (60 mg daily) and hydroxychloroquine (200 mg twice daily). During the first week of treatment, she lost 4 kg of body weight, the leg swelling and eosinophilia resolved, and the leg tenderness decreased considerably. During the next two weeks she lost an additional 6 kg. The hyperpigmentation persisted, but the skin tightness and tenderness resolved, and the prednisone dose was reduced to 10 mg daily.

Patient 3 was a 52-year-old woman with a history of chronic depression and a sleep disorder, treated with thiothixene (5 mg daily), doxepin (100 mg daily), temazepam (15 to 30 mg daily), and tryptophan (1.5 g daily). She had begun taking 1.0 g of tryptophan daily in early 1988. Eight months later, Bell's palsy developed. In July 1989, the daily dose of tryptophan was increased to 1.5 g. Six weeks later, she noted a pruritic erythematous rash on her trunk and arms, which occurred after drinking wine and lasted for one hour. In October, she had an infection of the upper respiratory tract accompanied by a low-grade fever, myalgia, and muscle cramps. A throat-swab culture was positive for streptococcus, and she took penicillin for 10 days. Her myalgia and cramps persisted, as did an intermittent urticarial rash, but she had no skin tightness or edema, and a blood smear showed 39 percent eosinophils.

The patient was first seen by us in November 1989, five days after having discontinued tryptophan therapy. Her muscle symptoms had improved, and her muscle strength was normal, but she had diffuse muscle tenderness and mild edema of the forearms. There was no evidence of Raynaud's phenomenon, sclerodactyly, telangiectasia, or cardiopulmonary disease. The eosinophil count was 3.26×10⁹ cells per liter, the serum creatine kinase level and liver function were normal, and the sedimentation rate was 10 mm per hour. A biopsy of the skin of the forearm revealed edema and inflammation involving the dermis and underlying fatty tissue.

She continued to have myalgia and muscle cramps, and when she was examined three weeks after the tryptophan therapy was discontinued, slight thickening of the skin of both forearms was noted. Her serum creatine kinase concentration was normal, but the aldolase concentration was elevated (13.0 IU per liter; normal range, 1.0 to 7.0). The absolute eosinophil count remained elevated (2.2×10⁹ cells per liter). Prednisone was prescribed in a dose of 20 mg daily. Within five days, her myalgia and muscle cramps had improved, she had lost 2 kg, and her eosinophil count was 0.33×10⁹ cells per liter. The dose of prednisone was increased to 60 mg daily, and the skin texture subsequently returned to normal; she lost an additional 5 kg, but a peripheral sensorimotor neuropathy developed.

Histopathologic Studies

Skin-biopsy specimens were available from all nine patients. Sections prepared by routine techniques were stained with hematoxylin and eosin, Masson's trichrome stain, chloroacetate esterase (Leder) stain for mast cells and neutrophil precursors, Giemsa stain, and Schiff's methylene blue stain for plasma cells.

Biochemical Studies

After giving informed consent, six patients and five age- and sex-matched normal subjects underwent studies of tryptophan metabolism according to a protocol approved by the institutional review board for human research of the Medical University of South Carolina. The subjects followed a diet providing 1800 kcal and 70 g of protein during the two study days. On the first day, a 24-hour urine sample was collected for the measurement of 5-hydroxyindoleacetic acid (5-HIAA) excretion (Nichols Institute, San Juan Capistrano, Calif.). On the second day, plasma samples were obtained before the oral administration of 1 g of tryptophan (Goldline Laboratories, Fort Lauderdale, Fla.) and at 30– to 60-minute intervals for eight hours afterward, and a second 24-hour urine sample was collected. The plasma samples were frozen at − 70°C before analysis. Tryptophan, L-kynurenine, and quinolinic acid were measured according to previously described methods.^{6 7 8 9} Statistical analysis was performed with an unpaired t-test after logarithmic transformation of the data.

Clinical Features

Table 1. Table 1. Characteristics of Patients with Scleroderma, Fasciitis, and Eosinophilia Associated with the Ingestion of Tryptophan.*

Six women and three men (mean age, 47 years; range, 30 to 66) were studied, including a husband and wife (Table 1). Symptoms developed in all subjects 1 to 18 months (mean±SD, 5.8±5.4) after tryptophan therapy was begun. The daily doses of tryptophan ranged from 1.0 to 3.0 g (mean, 2.0±0.7), and the mean cumulative dose was 307±209 g (range, 45 to 610 g) at the time of the onset of symptoms. Six patients were taking other medications when their symptoms appeared. Their occupations were diverse — homemaker, shop owner, and textile worker — and two were retired.

Table 2. Table 2. Clinical Features of Patients with Scleroderma, Fasciitis, and Eosinophilia Associated with the Ingestion of Tryptophan.

Figure 1. Figure 1. Skin Involvement in the Nine Patients with Localized Scleroderma, Fasciitis, and Eosinophilia.

The white areas denote normal skin, grey areas edematous and slightly tight skin, and black areas hidebound skin.

All nine patients had similar scleroderma-like abnormalities of the skin (Table 2). Their early symptoms included swelling of the skin, often accompanied by weight gain, urticaria, pain, and paresthesia, which evolved to become tight skin that was frequently pruritic and painful. Most patients had pitting edema, which had a woody consistency, and some had peau d'orange changes of the skin of the legs. The distribution of skin involvement was symmetric and followed a similar pattern in all patients (Fig. 1). There was a predilection for the extremities at the onset, particularly the legs; the trunk was often involved later. Only one patient had involvement of facial skin, and none had Raynaud's phenomenon or sclerodactyly. Flexion contractures occurred in the more severe cases, but no tendon friction rubs were detected. Microscopical examination of nail-fold capillaries was normal, indicating normal microvasculature in all patients.⁵ Five patients had myalgia and muscle cramps. Serum creatine kinase levels were normal; two patients had elevated serum aldolase levels. Two patients had paresthesia and compression neuropathy of the hands and feet, which resolved. Two other patients had peripheral sensorimotor neuropathy later in their illness, which persisted.

None of the patients had visceral-organ involvement characteristic of systemic scleroderma (Table 2). None had dysphagia, and barium esophagograms were normal in all five patients studied. Patient 1 had a transient pleural effusion and eosinophilic pneumonitis. Patient 3 had symptoms of an upper respiratory tract infection and a normal chest film early in the course of her illness. None of the other patients had respiratory symptoms or abnormal chest films or pulmonary function.

All the patients had eosinophilia. The percentage of eosinophils in the blood ranged from 4 to 52, and the absolute eosinophil counts ranged from 1.46×10⁹ to 10.80×10⁹ cells per liter (mean, 3.62±2.87× 10⁹). Other blood-cell counts were normal in all patients. Two patients had antinuclear antibodies. Antibodies to topoisomerase or centromere antigens were not detected. Most patients had normal sedimentation rates, and their serum albumin and globulin levels were low or moderately low, particularly during the early phase of their illness.

Treatment and Course

After tryptophan was discontinued, all but one patient (Patient 7) were treated with prednisone. Patient 7 had stopped taking tryptophan one month after the onset of her skin rash. Her skin lesions softened and her eosinophilia resolved during the subsequent 11 months. When she was examined by us, the skin of the anterior part of her legs was hyperpigmented and hidebound, the absolute eosinophil count was 0.06×10⁹ cells per liter, and a skin biopsy showed dermal inflammation and fibrosis without inflammation of the fascia. In the other patients, the initiation of prednisone therapy, in doses of 30 to 60 mg daily, resulted in rapid decreases in the eosinophil counts to normal, prompt and dramatic weight loss (up to 12 kg in one week in several patients), and marked improvement in cutaneous signs and symptoms. In only two patients (Patients 5 and 8), however, did the skin abnormalities completely disappear, permitting the discontinuation of prednisone. Myalgia generally resolved promptly after prednisone therapy was instituted, but two patients (Patients 3 and 9) had persistent muscle weakness and peripheral neuropathy despite continued therapy.

Histopathological Features

Table 3. Table 3. Histopathological Findings in Patients with Scleroderma, Fasciitis, and Eosinophilia Associated with the Ingestion of Tryptophan.*

Figure 2. Figure 2. Skin-Biopsy Specimen from Patient 1 (Hematoxylin and Eosin, ×40).

The epidermis (top) is normal, whereas the dermis shows extensive collagen deposition and collagen-encased ectatic sweat glands.

All nine patients had changes in the dermis and subcutaneous tissue typical of scleroderma (Table 3). Although the biopsy specimens were obtained at various intervals after the onset of the skin lesions, all showed infiltration of inflammatory cells, indicative of ongoing disease activity and extensive collagen deposition and collagen-encased ectatic sweat glands (Fig. 2). The infiltrates were located around the small blood vessels, the remaining hair follicles and eccrine glands, and the nerves. An inflammatory-cell infiltrate was also found in and adjacent to the subcutaneous septa in all patients; it was composed mainly of lymphocytes and monocytes, but it included occasional plasma cells in all patients and eosinophils in seven patients. Increased numbers of mast cells were noted in Patients 5 and 6, but few extracellular granules were seen. No evidence of vasculitis was present in any of the biopsy specimens.

The subcutaneous connective tissue was abnormal in all patients. There was replacement of fat cells by collagen. The septa were thickened and fibrotic.

Figure 3. Figure 3. Fascia from Patient 5 (Hematoxylin and Eosin).

In Panel A, the fascia is thickened, and there is a dense inflammatory-cell infiltrate in the subcutaneous tissue just above it (×80). In Panel B, a higher magnification reveals that the inflammatory-cell infiltrate is composed of plasma cells, lymphocytes, macrophages, and eosinophils (×250).

The biopsy specimens from seven patients included a portion of the superficial fascia (Fig. 3); six contained inflammatory infiltrates ranging from mild to severe. Eosinophils were prominent in the fascia only in Patient 5.

Figure 4. Figure 4. Fascia and Muscle from Patient 6 (Hematoxylin and Eosin, ×50).

The fascia is thickened and contains an inflammatory-cell infiltrate composed of lymphocytes, macrophages, and plasma cells. The inflammatory-cell infiltrate extends into the underlying skeletal muscle, where many cells are seen abutting the skeletal-muscle bundles.

The biopsy specimens from four patients included skeletal muscle. In Patient 1, there was no inflammation within the skeletal muscle, nor was there evidence of degeneration or regeneration of muscle fibers. The biopsy specimens from Patients 3, 6, and 7 showed a moderately dense infiltrate of lymphocytes, macrophages, and plasma cells among the muscle fascicles just beneath the inflamed fascia, but no degenerating muscle cells (Fig. 4).

Biochemical Studies

Table 4. Table 4. Plasma Tryptophan and Its Metabolites in Patients with Scleroderma, Fasciitis, and Eosinophilia and in Normal Subjects.*

Six patients and five normal subjects underwent studies of tryptophan metabolism before and after a 1.0-g oral dose of tryptophan, and the plasma of an additional patient was analyzed for tryptophan metabolites. These studies were done at various times (days to months) after the withdrawal of tryptophan. The mean plasma tryptophan level was similar in the six patients and the five normal subjects (48.4± 16.6 vs. 42.0±11.2 μmol per liter; Table 4). The patients and normal subjects had similar increases in plasma levels of tryptophan; the peak values occurred 30 to 90 minutes after the oral administration of tryptophan. The mean basal 24-hour urinary excretion of 5-HIAA was similar in the patients and the normal subjects (17.9±7.0 vs. 18.8±3.7 μmol per day). The values were also similar on the day after tryptophan administration (21.2±8.7 vs. 19.4±5.8 μmol per day). Two of the three untreated patients (Patients 2 and 8) studied in the active, eosinophilic phase of their illness had substantial increases in urinary 5-HIAA excretion after the oral administration of tryptophan (Table 4).

The base-line plasma level of L-kynurenine was significantly higher in the seven patients than in the five normal subjects (5.35±3.56 vs. 1.65±0.52 μmol per liter; P<0.05) (Table 4). The four patients studied during the active phase of their illness, after the withdrawal of tryptophan but before the start of prednisone therapy, had plasma L-kynurenine levels that were significantly higher than those in the three patients studied later (two of whom were taking prednisone) or the normal subjects (8.06±1.53 vs. 1.73±0.35 and 1.65±0.52 μmol per liter; P<0.001). When normalized for base-line L-kynurenine values, L-kynurenine clearance was similar in the patients and normal subjects (data not shown). The plasma quinolinic acid levels (Table 4) paralleled those of L-kynurenine, suggesting that the elevated L-kynurenine values were not due to a block in the kynurenine pathway.

We describe a syndrome in nine patients taking tryptophan that was characterized by rash and edema of the extremities, frequently followed by scleroderma-like skin changes, myalgia and muscle weakness, and peripheral neuropathy and eosinophilia, but with little evidence of visceral involvement. The withdrawal of tryptophan had little or no immediate benefit. All but one patient received prednisone, but only two had complete resolution of symptoms during a follow-up period of up to one year. The patients whose illness was in the active phase had elevated plasma levels of L-kynurenine and often had exaggerated responses of plasma L-kynurenine and urinary 5-HIAA to an oral dose of tryptophan. We described a similar syndrome in 1980 in a patient treated with the tryptophan metabolite L-5-hydroxytryptophan and the aromatic L-amino acid decarboxylase inhibitor carbidopa.⁴ Like the patients in this series, our earlier patient also had elevated basal levels of plasma L-kynurenine and exaggerated responses of plasma L-kynurenine and urinary 5-HIAA to tryptophan. Taken together, these and other data substantiate an association between fibrosing illnesses and abnormalities of both the serotonin and the kynurenine pathways of tryptophan metabolism.^{10 11 12 13 14 15} Recent in vitro studies indicate that tryptophan and its metabolites, including L-5-hydroxytryptophan, kynurenine, 5-hydroxykynurenamine, and quinolinic acid, could induce many of the major features of the syndrome, such as fibroblast proliferation, vascular permeability, and neurotoxicity.^{16 17 18 19 20}

Tryptophan is an essential amino acid that has been used to treat insomnia in doses ranging from 1 to 5 g daily.²¹ Although it has been presumed to have no undesirable side effects, some investigators have speculated that it might be neurotoxic in certain predisposed persons.²⁰ Our patients were taking from 1 to 3 g daily, either at the behest of their physicians or on their own initiative, to treat sleep disturbances or depression or as part of dietary supplementation to treat obesity. Within 1 to 18 months after the initiation of treatment, signs and symptoms of scleroderma with fasciitis and eosinophilia developed in these patients. Muscle involvement was common, and two patients had peripheral sensorimotor neuropathy. One patient had transient eosinophilic pneumonitis, but none had typical features of scleroderma.

The clinical and histopathological findings, particularly during the acute phase of our patients' illness, are consistent with, but not identical to, those of eosinophilic fasciitis.^{22 23 24} Since the nosology of scleroderma is descriptive and imprecise, the illness in our patients might be classified as localized scleroderma, generalized morphea, morphea profunda, or eosinophilic fasciitis.^{22 23 24 25 26} Despite the nosologic confusion, the illness associated with tryptophan ingestion clearly belongs within the spectrum of scleroderma disorders. We have chosen to designate it descriptively as scleroderma with fasciitis and eosinophilia. It remains to be determined whether all of the recently described cases of the eosinophilia—myalgia syndrome associated with tryptophan ingestion belong within the same spectrum of disease.^{1 , 2}

The most pressing issue in preventing further occurrences of this syndrome is the determination of whether the syndrome is related to a contaminant of the tryptophan preparation or to the ingestion of tryptophan itself in certain predisposed persons. The identification of predisposing host factors would then be critical in defining patients at risk for the syndrome. Of particular relevance in distinguishing between underlying susceptibility and a contaminant is our previous report,⁴ and the reports of others,^{27 , 28} of the same syndrome occurring in patients treated with L-5-hydroxytryptophan and tryptophan (with or without the aromatic L-amino acid decarboxylase inhibitor carbidopa). The occurrence of the same syndrome in patients receiving two different tryptophan metabolites many years apart lessens the likelihood that the current epidemic is related to chance contamination of the tryptophan preparations. Rather, it suggests either that a contaminant, if present, is inherent in a production process common to the two agents or that the syndrome is not related to the common contamination of the preparations but is associated with underlying host susceptibility. The biochemical findings in the patient with scleroderma related to the ingestion of L-5-hydroxytryptophan and carbidopa suggested that differences in tryptophan metabolism may be associated with the development of fibrosing syndromes in certain persons. Our current findings of similar alterations in tryptophan metabolism in patients and in age- and sex-matched normal subjects suggest that underlying factors may also have a role in the pathogenesis of the tryptophan-related syndrome.

The pattern of elevated plasma L-kynurenine and quinolinic acid concentrations in patients in the active phase of their illness is consistent with increased activity of the rate-limiting enzyme in the kynurenine pathway, indoleamine-2,3-dioxygenase. Interferon gamma and endotoxin stimulate the activity of this enzyme.^{29 30 31 32} Thus, the biochemical differences in plasma L-kynurenine and quinolinic acid levels and urinary 5-HIAA excretion between the patients and the normal subjects could be due to stimulation by inflammatory mediators such as interferon gamma or proinflammatory agents such as endotoxin.^{9 , 32} Although the pattern of tryptophan metabolites in these patients is consistent with the response to endotoxin and other proinflammatory agents, there is no direct evidence that exposure to endotoxin initiated the syndrome in them.

One of us has recently shown in rats that susceptibility to arthritis induced by the streptococcal cell wall (bacterial lipopolysaccharide and endotoxin) is related to the interruption of the loop by which inflammatory mediators signal the hypothalamic—pituitary—adrenal axis to restrain the intensity of the inflammatory response, through the potent anti-inflammatory effects of glucocorticoid.³³ In this regard, it is striking that six of these nine patients were either receiving drugs (benzodiazepines) known to inhibit the secretion of corticotropin-releasing hormone^{34 , 35} or had intrinsic insufficiency of the hypothalamic—pituitary—adrenal axis (Addison's disease). Furthermore, the patients who had the L-5-hydroxy-tryptophan—related illness were also receiving benzodiazepines.^{4 , 28 , 29} Although the use of benzodiazepines in these patients may reflect selection bias, in that patients taking tryptophan for insomnia might be more likely to take benzodiazepines for anxiety-related symptoms, not all the patients in whom the syndrome developed were taking the same psychotropic drug.

This confluence of factors suggests a working hypothesis for the pathogenesis of the scleroderma-like syndrome related to the ingestion of tryptophan. Although the precise environmental trigger of the syndrome is not known and our biochemical data do not rule out the possibility of a contaminated tryptophan preparation, the findings are consistent with activation of the enzyme indoleamine-2,3-dioxygenase by an agent such as endotoxin. It is conceivable, therefore, that the syndrome developed because of a confluence of factors: exposure to an environmental inflammatory trigger such as endotoxin, simultaneous impairment of hypothalamic—pituitary—adrenal function, and concurrent ingestion of tryptophan, all of which lead to the specific features of the syndrome, including edema, inflammation, and fibrosis.