American Experience With Low-Dose Thalidomide Therapy for Severe Cutaneous Lupus Erythematosus FREE

THALIDOMIDE (α-N-phthalylglutamic-acid-imide), a glutamic acid derivative, was first synthesized in 1956 by Kunz and his coworkers in West Germany.^{1 - 2} Thalidomide became widely used as a sleeping aid but was removed from the market after teratogenic and neurologic effects were established.^{3 - 12} Subsequently, thalidomide became available only for experimental purposes. There was renewed interest in the clinical uses of thalidomide after Sheskin¹³ reported in 1965 the effectiveness of thalidomide in treating erythema nodosum leprosum. Sheskin, after giving thalidomide to patients with leprosy for sedative purposes, made a serendipitous discovery that the treatment resulted in immediate improvement in symptoms, and that the patients' cutaneous erythematous lesions resolved in as little as 3 days. His findings were confirmed by other reports^{14 - 16} and by his follow-up study.¹⁷

Recently, thalidomide therapy has been approved by the Food and Drug Administration for the treatment of erythema nodosum leprosum. However, it remains an experimental treatment for a variety of inflammatory skin disorders.¹⁸ Among those reported included actinic prurigo,^{19 - 20} orogenital ulcerations,^{21 - 23} Weber-Christian disease,²⁴ graft-vs-host disease,^{25 - 28} and chronic cutaneous lupus erythematosus (CCLE).^{29 - 30}

Many European studies have documented that thalidomide is an effective treatment for various inflammatory dermatoses, such as cutaneous lupus erythematosus (CLE), that are unresponsive to standard therapy. Knop et al,³¹ in a study of 60 cases of CCLE, reported that 90% of the patients had a complete or marked response to treatment with thalidomide, 400 mg/d, and that 71% had relapses when thalidomide was withdrawn. Stevens et al³² had similar findings with thalidomide, 50 to 100 mg/d, with various adjustments, if any, in the doses of other oral medications already being used to treat CLE. Improvements were observed within 2 weeks of treatment, and maximum benefits were achieved within 16 weeks. Another study involving 11 patients reported that a maintenance dose of 25 to 50 mg every night or every other night was necessary to prevent relapse.³³ Other studies were also published confirming the effectiveness of thalidomide, 100 to 400 mg/d, in the cutaneous treatment of various forms of lupus erythematosus (LE).^{34 - 37} The reported adverse effects in these studies included paresthesia, drowsiness, abdominal disturbances, dry mouth, urticaria, rash, mood changes, circulatory changes, amenorrhea, and edema. Of these, paresthesia, sedation, and constipation were the most common.

Although thalidomide is well recognized as an effective agent for CLE through the studies done mainly in Europe, it is rarely used to treat this disorder in the United States because of its risks and because it is difficult to obtain. For these reasons, there are no randomized trials that have proved the efficacy of thalidomide as a second-line drug in CLE. Herein, we report our experience in treating 7 patients with CLE during 9 years. Our experience confirms its remarkable efficacy and surprising safety as monotherapy.

We reviewed the medical records of 7 women with disfiguring CLE who were treated with thalidomide in a 100-mg/d starting dose after exhausting the other therapeutic options, including glucocorticosteroids, antimalarials, immunosuppressants, dapsone, clofazimine, and retinoids. These patients continue to receive thalidomide therapy. Candidates for thalidomide therapy were selected by the presence of severe CLE that had previously been treated with first- and second-line therapies.³⁸ This treatment was offered on a compassionate basis by the Department of Dermatology, University of Rochester, Rochester, NY, for patients who had been referred by community dermatologists. Table 1 lists their conditions and Table 2 describes the therapeutic modalities that failed. Candidates were further selected by their ability to understand the risks and benefits of thalidomide therapy. Further requirements included a willingness to comply with the monitoring requirements, and to sign an informed consent (reviewed and approved by the University of Rochester Research Subjects Review Board). After a candidate for thalidomide therapy was identified, a single-patient investigational new drug application was submitted to the Food and Drug Administration. After this application was submitted and approved (usually 3-6 months after submission), the Food and Drug Administration issued an investigational new drug number (see Table 3). Contraception was initiated in women of childbearing age. A prescription with payment and IND number were mailed to the supplier of thalidomide. This drug was then mailed to the prescribing physician, who dispensed the thalidomide to the patient.

All 7 patients were treated according to the published guidelines for thalidomide use.³⁹ All other systemic medications for controlling CLE were discontinued. The patients returned every 3 months for follow-up, and the efficacy and adverse effects of their thalidomide treatment were documented. Nerve conduction studies, complete blood cell counts, 12-factor automated blood chemistry studies, and thyroid studies (free thyroxine, thyrotropin) were done every 6 months or sooner as needed for baseline comparison. Any change in blood cell count or results of serum chemistry studies was followed up closely to determine if it was thalidomide related. All patients were examined at baseline with a complete neurologic examination, with a focus on clinical assessment of touch, scratch, vibratory, and position sensations. These examinations were repeated at 6-month intervals during thalidomide therapy. Nerve conduction studies were done in all patients at baseline and at 6-month intervals to detect evidence of drug-induced sensory or motor loss in the extremities. The nerve conduction studies included unilateral examination of the arm and leg. Standard techniques of percutaneous stimulation and surface recording were used. Motor conduction velocity, distal motor latency, and the amplitude of the compound muscle action potential were measured from the common peroneal, tibial, ulnar, and/or median motor nerves. Sensory conduction velocity, sensory distal latency, and the amplitude (baseline to negative peak) of the sensory nerve action potential (SNAP) were obtained from the sural, superficial peroneal, ulnar, and/or median sensory nerves. The patients understood that they must stop taking thalidomide should they begin to experience paresthesia. All patients continued sun avoidance and use of sunscreen.

Table 3 summarizes the course of each patient's treatments; Table 4 describes their reported adverse effects. Six of the 7 patients responded to thalidomide treatment. One patient dropped out after 1 month of thalidomide treatment with concerns about the possible adverse effect of neuropathy (this patient had not experienced any adverse effects at the time of discontinuing thalidomide). Of the 6 patients who responded to treatment, 2 had complete resolution of their cutaneous lesions by the end of the first year of treatment. The other 4 showed marked (80%) improvement in their skin, with scarring and hyperpigmentation remaining in the scalp or chest. One patient had hair regrowth after 2 months of treatment, and this regrowth reached a plateau by the sixth month of treatment. Figure 1 summarizes the effects of thalidomide therapy on the cutaneous lesions in patients 3 through 6. Figure 2 depicts the effects of thalidomide therapy on the histological changes of CCLE in patient 5.

Table 4 lists the adverse effects reported to us by the patients. The most common adverse effects reported were paresthesia, sedation, constipation, and weight gain. Sedation and constipation were for the most part well tolerated by the patients. Most became accustomed to the drug's sedative effect, and 2 patients reported insomnia when not taking thalidomide. Constipation was tolerable after dietary changes and taking laxatives as needed. No treatment was recommended for weight gain, since this adverse effect was transient and not distressing to any of the patients. There were no complaints of decreased libido.

Nerve conduction studies did not demonstrate abnormalities that specifically related to treatment with thalidomide. The motor nerve conduction data were all within the normal range for distal motor latency (peroneal, <5.8 milliseconds; tibial, <6.2 milliseconds; ulnar, <4.0 milliseconds; median, <4.2 milliseconds), for compound muscle action potential amplitudes (peroneal, >2.2 mV; tibial, >6.0 mV; ulnar, >4.0 mV; median, >6.0 mV), and for motor conduction velocities (peroneal, >40 m/s; tibial, >38 m/s; ulnar, >52 m/s; median, >50 m/s). Sensory nerve conduction was also within the normal range with the exception of 1 patient. This patient developed a decline and loss of the SNAP signals from her left sural and and left superficial peroneal nerves in association with clinical and imaging findings of a compressive left L5-S1 radiculopathy. Otherwise, the sensory nerve conduction data for the patients were within the normal ranges for the amplitudes of SNAPs (sural, >6 µV; superficial peroneal, >6 µV; ulnar, >17 µV; median, >20 µV) and for the sensory conduction velocities (sural, >40 m/s; superficial peroneal, >40 m/s; ulnar, >49 m/s; median, >48 m/s). Amplitudes of the SNAPs varied in the same nerve on different studies, but there was no consistent trend in the values to indicate a significant decline in amplitudes. Coexisting clinical conditions rather than the treatment with thalidomide caused neurologic symptoms.

Four patients reported experiencing paresthesia, but sensory nerve conduction studies did not disclose any drug-induced abnormalities. None of the 4 cases could be linked to thalidomide by neurologic evaluation and nerve conduction studies (see Table 4). In 1 patient, perioral paresthesia resolved when thalidomide treatment was withdrawn for 2 weeks and did not recur when treatment was restarted. The second patient had improvement in neurologic symptoms while still receiving thalidomide; nerve conduction studies showed abnormalities consistent with her underlying connective tissue disease but not with thalidomide neuropathy. The third patient was found to have paresthesia caused by a C6 compression of a nerve root. The last patient recently developed paresthesia after 9 years of thalidomide treatment. Nerve conduction studies did not show a thalidomide neuropathy.

It is of interest that 2 patients reported similar adverse effects of intermittent "shaking" episodes since starting thalidomide treatment. These shaking episodes were described as being similar to tremors and chills. They were preceded by thoughts of impending shakes by the whole body. These symptoms initially occurred in the mornings, progressed to occur all day, and then gradually resolved. No hematologic, biochemical, or nerve conduction abnormalities were found that could explain this phenomenon. There were no abnormalities in body temperature associated with these shaking episodes. In 1 patient, this symptom resolved while the patient was still receiving thalidomide. Since complete neurologic examinations did not show any abnormalities, magnetic resonance imaging of the brain was not performed.

No significant changes in blood cell count were noted in any of the patients after initiation of thalidomide treatment. One patient had a history of slightly depressed white blood cell count that did not change after thalidomide was started. No electrolyte disturbances were noted, and the serum urea nitrogen and serum creatinine levels in the patients with systemic LE remained stable at baseline. None of these patients developed abnormalities in thyroid function tests (data not shown).

In patients who had circulating autoantibodies at baseline (patients 1, 3, 5, and 6), thalidomide therapy did not decrease these circulating antibodies, despite its beneficial effects on their dermatologic disease (data not shown). This suggests that the thalidomide may be working by an anti-inflammatory mechanism, rather than the immunomodulation of a pathogenic immune response.

Cutaneous lesions secondary to systemic LE, subacute CLE, or CCLE in the 6 patients who took thalidomide, 100 mg/d, for more than 1 month all showed marked improvement. The average response time was 2.2 ± 0.8 months. The patients' conditions remained stable even after discontinuing other treatment modalities, including oral, injected, or topical glucocorticosteroids, and other immunosuppressive or modulating pharmacologic agents (Table 2). In addition to the resolution of the clinical signs and symptoms associated with CLE (Figure 1), thalidomide also resulted in the resolution of the histological changes associated with CLE (Figure 2). The patients with systemic LE have not shown signs of their systemic disease since starting thalidomide and discontinuing other treatments. The use of sunscreens and sun avoidance remains a must, since most of these patients have histories of photosensitivity.

THERE IS a general lack of agreement on the effective starting dose and treatment schedule of thalidomide. In our experience, 100 mg daily in the evening is an adequate starting dose. We have been successful in tapering the dose of thalidomide in 3 patients (patients 1, 5, and 6) (from daily dosing to alternate-day or every-third-day dosing with 100 mg of thalidomide). In the other patients (patients 2 through 4), our attempts to taper the dose of thalidomide resulted in a prompt exacerbation of their condition, which necessitated the continuation of a daily dosing regimen.

Complete discontinuation of thalidomide therapy has not been possible in any of our patients, because of relapses in their dermatologic disease. This is in contrast to what has been suggested by other clinicians, that the dose be gradually decreased until treatment can be discontinued altogether.³³ The induction of long-term remission by thalidomide has not been reported, and this has been our experience as well. This suggests that a maintenance dose is necessary for long-term control of disease.

One of the patients with CCLE (patient 1) was initally treated with thalidomide provided by a Hansen disease treatment center (Table 3). This source of thalidomide was interrupted because of a drug shortage, which necessitated obtaining thalidomide from a different supplier (Pediatric Pharmaceuticals, Iselin, NJ). Although we have no information about the relative bioavailabilities of thalidomide from the 2 suppliers, this patient maintained an excellent clinical response to thalidomide provided by a different supplier. There are 3 American pharmaceutical companies that manufacture and distribute thalidomide, but relative bioavailabilities of thalidomide from these sources is not known. This is a potentially important issue that warrants study.

The mechanism of action of thalidomide remains unknown, but the drug is known to possess certain anti-inflammatory and immunomodulatory properties. The actions of thalidomide include the inhibition of monocyte tumor necrosis factor α production,⁴⁰ inhibition of polymorphonuclear leukocyte chemotaxis⁴¹ and phagocytosis,⁴² inhibition of lymphocyte proliferation in response to mitogenic or allogenic stimulation,⁴³ inhibition of helper T-cell type 1 cytokine production while enhancing helper T-cell type 2 cytokine production,⁴⁴ and inhibition of angiogenesis.⁴⁵ One or more of these mechanisms may explain thalidomide's effectiveness in controlling the cutaneous manifestations of LE.

The most serious adverse effects of thalidomide treatment are teratogenicity and neuropathy. Therefore, strict contraceptive measures and close monitoring of neurologic symptoms are necessary. In our patients, no serious adverse effects have been observed, even in those treated for longer than 1 year, and discontinuation of treatment has not been necessary. One patient elected to discontinue because of concerns about the potential adverse effect of neuropathy.

Some studies of thalidomide neuropathies associated with the treatment of dermatologic diseases have suggested that the occurrence of neuropathy may be related to cumulative thalidomide dose (emerging after >40- to 50-g total dose), whereas other reports have indicated that the cumulative dose was not important.^{12 ,46} There is a suggestion that host-related factors (female sex and advanced age) may be critical risk factors in patients who develop neuropathy in the first few months of treatment.^{5 ,47 - 48} In our series of 7 female patients, 4 have taken a cumulative dose of thalidomide greater than 50 g without a detectable neuropathy. Three of these patients were younger than 50 years. It is likely that the small sample size, the relatively young age of our patients, the brief duration of thalidomide therapy, the low doses used, and an absence of preexisting neurologic disease were all factors responsible for a lack of thalidomide neurotoxic effects in this series.

Careful serial clinical examinations with detailed testing of peripheral nerve function appear to offer the most suitable approach to detect early signs related to thalidomide-induced neuropathy. The occurrence of neuropathic complaints may vary depending on the disorder requiring thalidomide treatment, and genetic as well as other predisposing factors may influence the likelihood that neuropathy will develop.⁴⁹ The use of electrodiagnostic testing to identify thalidomide-induced neuropathy is also not as reliable as one might hope. The amplitudes of SNAP are quite variable on reproducibility testing.^{49 - 51} Some investigators have grouped the SNAP measurements to adjust to this variability in the method.⁴⁹ Most of the disorders that require thalidomide therapy have an increased likelihood of other causes of neuropathy, including compression neuropathy, such as carpal tunnel syndrome, which by itself can cause a decline in the SNAP. This increased tendency toward various types of neuropathy emphasizes the need for clinical examination and judgment to establish whether the symptoms are caused by toxic effects of thalidomide or another process. For example, one study reported the incidence of thalidomide neuropathy to be 21% to 50% in patients with various dermatoses, without significant difference in incidence between the different conditions.^{49 ,52} It is interesting that the sensory nerve evaluation in these reports used the SNAP amplitudes for sural and median nerves. The median nerve is particularly prone to compression neuropathy. Perhaps the variation in incidence of neuropathy observed in these previous reports partly results from factors other than toxic effects of thalidomide. Hafstrom⁴⁸ reported that decreased SNAPs with normal conduction velocities, a characteristic finding in drug-induced neuropathy, can also be seen in asymptomatic patients. In our patients who have not reported paresthesia, repeated nerve conduction studies have not shown any sensory or motor loss. Continued monitoring of patient 3, who has a history of carpal tunnel syndrome, is necessary, since underlying carpal tunnel syndrome may be a risk factor for thalidomide-induced peripheral neuropathy.⁵³ In summary, careful, serial neurologic evaluations of patients receiving thalidomide provides the single most effective means to detect the early development of neuropathy. Nerve conduction testing is a useful adjunct to the clinical assessment.

The minor adverse effects have been tolerable to the patients. We continue to monitor the symptom of shaking. This phenomenon has not been previously reported, and we have not been able to link it to thalidomide-induced biochemical, hematological, or neurologic abnormalities. This symptom has not been a dose-limiting toxic effect. In one patient, the symptom has resolved without interrupting treatment; in the other patient, this symptom has persisted but has not been dose limiting.

Overall, the patients who have shown improvement have been pleased with the results of thalidomide therapy. The patients have perceived a benefit in arresting the progression of their chronic, previously unrelenting dermatologic disease. The minor adverse effects of thalidomide have been tolerable in all patients, since all of them have been afflicted with their dermatoses for many years (Table 1).

Various studies, mainly in Europe, have reported the safety and effectiveness of various doses of thalidomide in the treatment of the CLE. In our experience, a low starting dose of thalidomide has been a safe and effective monotherapy for CLE after traditional therapeutic options have been exhausted. Minor adverse effects are common but tolerable. The dispensing of thalidomide is tightly regulated in the United States, given the drug's infamous history. Even so, thalidomide is a promising treatment option for clinicians to consider, but caution and familiarity with the prescribing guidelines cannot be overemphasized (Table 5).

Accepted for publication April 21, 1999.

This study was supported in part by grant 1R01AH/OH4108-01 from the National Institutes of Health, Bethesda, Md (Dr Gaspari), and by a grant from the Lupus Foundation, Genessee Valley Chapter, Rochester, NY (Mr Duong).

Reprints: Anthony A. Gaspari, MD, Department of Dermatology and Cancer Center, University of Rochester School of Medicine and Dentistry, Rochester, NY 14642 (e-mail: anthony_gaspari@urmc.rochester.edu).