Patient-Reported Outcomes and Health-Care Resource Utilization in Patients with Psoriasis Treated with Etanercept: Continuous versus Interrupted Treatment

Authors


Seth R. Stevens, Amgen Inc., One Amgen Center Drive, MS 27-1-D, Thousand Oaks, CA 91320, USA. E-mail: sstevens@amgen.com

ABSTRACT

Objective:  The 24-week Etanercept Assessment of Safety and Effectiveness (EASE) study evaluated the effectiveness and tolerability of continuous versus interrupted etanercept treatment in patients with moderate to severe plaque psoriasis. The objective of this analysis was to assess patient-reported outcomes (PROs) and health-care resource utilization (HRU) data from the EASE study.

Methods:  Patients received open-label etanercept 50 mg twice weekly for 12 weeks and then received either continued or interrupted (single round of discontinuation and re-treatment with etanercept) etanercept 50 mg once weekly for the second 12 weeks. PROs included the following: 1) the patient global assessments of psoriasis, joint pain, and itching scores; 2) the Dermatology Life Quality Index; 3) the Medical Outcomes Study Short Form 36 vitality domain; 4) the Beck Depression Inventory; 5) the European Quality-of-Life Group Feeling Thermometer; and 6) a patient satisfactionsurvey. HRU was evaluated using the Economic Implications of Psoriasis patient questionnaire.

Results:  Continuous treatment with etanercept 50 mg twice weekly for 12 weeks followed by 50 mg once weekly for 12 weeks produced sustained and clinically important improvements in PROs and reductions in HRU. Reductions in some outcome measures after treatment discontinuation at week 12 were observed in the interrupted group; however, most changes did not revert to baseline levels, consistent with some residual clinical effect, and re-treatment produced improvements similar to week 12 levels.

Conclusions:  Continuous etanercept treatment provided greater sustained improvements in PROs than interrupted therapy; however, interrupting etanercept therapy, if needed, has predictable and manageable effects.

Introduction

Psoriasis is a chronic, debilitating inflammatory disease [1] that affects approximately 2% of the US population [2] and has a profound effect on the lives of patients and their family members. About 40% of patients with psoriasis indicate that their condition is a problem in their everyday lives [2]. Patients with more severe disease and those with visible psoriatic plaques report greater disability and lower health-related quality of life than those with less severe disease or those with plaques on less visible areas of the body [2–4].

Psoriasis can have markedly negative consequences on the physical and psychosocial functioning of those affected, with patients often experiencing embarrassment, self-consciousness, and negative body image as a result of their condition [5]. Depression [6] and even suicidal ideation [7] can also be important issues of concern in patients with psoriasis. Furthermore, the unique disabilities and emotional consequences associated with this disease can affect work and social activities [5,8,9]. In addition to the psychosocial aspects of psoriasis, limited cost data suggest that annual health-care resource utilization (HRU) for this disease is considerable and represents a significant burden to patients and society [10].

Long-term treatment strategies are required to effectively manage psoriasis [1]. Treatment that can provide disease suppression or remission may offer patients relief from the burden of psoriasis. The fully human soluble tumor necrosis factor (TNF) receptor––FC fusion protein etanercept––is indicated for moderate to severe plaque psoriasis in the United States, Canada, and Europe, and has been shown to be effective in relieving the skin symptoms and in improving health-related quality of life in patients who have this disease [11–13].

Most clinical studies of psoriasis therapies have evaluated short-term efficacy with continuous treatment. In the real world, however, there are various life circumstances (e.g., interruption in insurance coverage or preparation for surgery) that may require intermittent use of therapy. Therefore, it is important to determine the impact of intermittent use of agents for psoriasis not only in terms of clinical efficacy and tolerability, but also on patient-reported outcomes (PROs) and HRU as well. The Etanercept Assessment of Safety and Effectiveness (EASE) study evaluated the clinical efficacy and tolerability of continuous versus interrupted etanercept treatment in patients with moderate to severe plaque psoriasis. The clinical results from the EASE study have been reported elsewhere [14]. This report presents PROs and HRU data from the EASE study, with the hypothesis being that continuous treatment provides greater benefits than interrupted treatment, but that some level of benefit is maintained during interrupted therapy and benefit is regained with reinstitution of therapy.

Methods

The 24-week EASE study was a multicenter, randomized, open-label, phase 3b study of patients with psoriasis from 325 sites. The methodology for the EASE study has been described in detail previously [14]. Patients were randomly assigned to one of two groups: a continuous treatment group and an interrupted treatment group as outlined in Figure 1. After treatment with subcutaneous (SC) etanercept 50 mg twice weekly (BIW) for the first 12 weeks, patients in the continuous group received uninterrupted SC etanercept 50 mg once weekly (QW) through week 24. In the interrupted group, responders at week 12, defined as those achieving a physician's global assessment (PGA) of psoriasis score of ≤2 (0 = clear, 5 = severe) and improvement from baseline, discontinued therapy. Upon relapse (i.e., loss of PGA responder status), treatment with SC etanercept 50 mg QW was restarted at either week 16 or 20 and continued through week 24. Nonresponders in the interrupted group at week 12 received SC etanercept 50 mg QW from weeks 13 to 24, the same treatment received by those in the continuous treatment group.

Figure 1.

Flow diagram of patient disposition. BIW, twice weekly; ETN, etanercept; QW, once weekly.

Patient-reported outcomes included the following: 1) the patient global assessments (PtGA) of psoriasis, joint pain, and itching scores, in which patients were asked to rate these aspects of their psoriasis on a scale from 0 (good or none) to 5 (severe); 2) the Dermatology Life Quality Index (DLQI), a validated 10-item outcome questionnaire designed to assess the patient's health-related quality of life [15], with total score ranging from 0 to 30 points and higher scores indicating worse outcomes; 3) the Medical Outcomes Study Short Form 36 (SF-36) vitality domain, a 4-item self-administered questionnaire designed to assess the patient's perception of fatigue [16,17], with scores ranging from 0 to 100 and lower scores indicating more fatigue; 4) the Beck Depression Inventory (BDI), a 21-item measure of mental status designed to assess patients' symptoms of depression, with scores ranging from 0 to 63 and higher scores indicating more symptoms of depression; 5) the European Quality-of-Life Group Feeling Thermometer (EuroQoL-FT), a visual analog scale that assesses the patient's health status, with scores ranging from 0 (worst imaginable health state) to 100 (best imaginable health state); and 6) a patient satisfaction survey, a 10-item questionnaire designed to assess patients' satisfaction with their psoriasis treatment overall and for specific items (scaling, itching, redness, tightness, bleeding, burning, fatigue, work/school, and social activities). The hypothesis was that interrupted treatment would have lower benefits on PRO measures than continuous etanercept treatment at week 24.

Health-care resource utilization was evaluated using the Economic Implications of Psoriasis patient questionnaire, which assesses the impact of psoriasis from an economic perspective, including health-related outcomes (hospitalizations, emergency department or urgent care clinic visits, and nondermatologist physician visits) and work-related outcomes (employment, job responsibility changes because of psoriasis, and sick days from work).

Data on PtGA of psoriasis, joint pain, and pruritus, and on the DLQI, were collected at baseline and weeks 2, 4, 12, 16, 20, and 24. The BDI, EuroQoL-FT, patient satisfaction survey, and Economic Implications of Psoriasis patient questionnaire were assessed at baseline and weeks 12 and 24. The SF-36 vitality domain questionnaire was assessed at baseline and weeks 2, 4, 12, and 24.

Analyses were performed on all patients who received at least one dose of study drug using the last-observation-carried-forward method to impute missing data. Absolute improvements and the percentage improvements from baseline by visit in PROs were summarized descriptively. In addition, 95% confidence intervals (CIs) were calculated by treatment group at all postbaseline visits for the efficacy end points. The proportions of patients in the continuous and interrupted arms were compared using a chi-square test. For the economic implications of psoriasis outcome measures, P-values for weeks 12 and 24 results relative to baseline were calculated using McNemar's test. The following categories of HRU were collected: number of days hospitalized in the last 3 months, number of visits to an emergency room or urgent care clinic in the last 3 months, and number of visits to a physician's office other than a dermatologist in the last 3 months. HRU data were skewed and were not normally distributed; therefore, 95% bootstrap bias corrected and accelerated CIs were reported (bootstrap sample: 1000). P-values were based on the Wilcoxon signed rank test comparing week 24 to baseline.

A post hoc analysis for PtGA and DLQI was performed on patients in the interrupted group who discontinued treatment at week 12 but restarted etanercept treatment before week 24. This analysis was performed to account for and to exclude patients in the interrupted group who actually continued treatment after week 12 (i.e., those who did not achieve a PGA responder status). A second hypothesis was generated post hoc, which postulated that among PGA responders who discontinued and subsequently resumed treatment, PROs would worsen from the discontinuation visit (week 12) to the re-treatment visit (week 16 or 20) and revert back to baseline levels. We were not able to assess the impact of discontinuation and re-treatment on BDI, SF-36 vitality subscale, or EuroQoL-FT because these outcomes were not assessed at the re-treatment initiation visit.

Results

Patients

There were 1272 patients in the continuous therapy group and 1274 in the interrupted therapy group who were randomized and received at least one dose of etanercept (N = 2546). Eighty-eight percent of patients in the continuous group and 85% of patients in the interrupted group completed the study. As per the study design, not all patients in the interrupted group discontinued therapy at week 12. Specifically, 357 patients (28%) in the interrupted group were PGA nonresponders at week 12 and continued etanercept therapy after the initial 12 weeks for the duration of the study (i.e., through week 24). Four-hundred eighty-eight patients discontinued and then restarted treatment with etanercept (Fig. 1). Demographic characteristics were well balanced between the two treatment groups (Table 1).

Table 1.  Baseline demographics and disease characteristics
CharacteristicsContinuous groupInterrupted group
Number of patients12721274
Mean (SD) age (year)45.8 (13.6)44.9 (13.6)
Men (%)6362
White (%)8486
Median (range) weight (kg)87.4 (41.4–218.3)87.8 (43.5–205.2)
Median (range) height (cm)173 (130–203)173 (127–203)
Median (range) duration of psoriasis (year)16.3 (0.0–63.0)16.4 (0.0–74.5)
Median (range) affected body surface area (%)20.0 (10.0–100.0)20.0 (10.0–100.0)
Physician global assessment of psoriasis, n (%)
 Clear0 (0)0 (0)
 Almost clear18 (1.4)16 (1.3)
 Mild191 (15.0)223 (17.5)
 Moderate683 (53.7)657 (51.6)
 Marked313 (24.6)313 (24.6)
 Severe67 (5.3)65 (5.1)
Patients with at least one prior systemic psoriasis therapy (%)4951

Efficacy and Safety

Efficacy and safety results are reported elsewhere [14]. Briefly, there were similar percentages of responders in the continuous and interrupted groups at 12 weeks (71.3% vs. 72.0%). Nevertheless, a significantly higher percentage of responders was observed at week 24 in the continuous group than in the interrupted group (71.0% vs. 59.5%; P < 0.0001). Nearly one-half (46.8%) of those who discontinued etanercept at week 12 remained off therapy through week 24, because they had not met the protocol-specified criteria for relapse based on PGA responder status at each of the subsequent evaluations. Etanercept was generally well tolerated in both treatment arms [14], with the safety results being consistent with those from placebo-controlled trials of etanercept [11–13].

PROs and HRU in the Continuous versus Interrupted Treatment Groups

PtGA of psoriasis, joint pain, and itching.  Mean PtGA of psoriasis scores were similar at baseline between the continuous and interrupted arms (3.9 vs. 4.0). The mean percentage improvement in PtGA of psoriasis scores was comparable between the continuous and interrupted groups at 12 weeks (Fig. 2a). These results were maintained from weeks 13 to 24 in the continuous treatment group, whereas the percentage improvement initially decreased in the interrupted group. A gradual improvement was observed starting at week 20 after patients had been allowed to reinitiate therapy.

Figure 2.

Mean percentage improvement from baseline in the patient global assessment of (a) psoriasis, (b) joint pain, and (c) itching. PtGA, patient global assessment; SEM, standard error of the mean.

Similar trends were observed in the PtGA of joint pain (Fig. 2b) and itching (Fig. 2c), in which mean scores were comparable between the two treatment arms at baseline and after 12 weeks of therapy. Similar to the PtGA of psoriasis, the percentage improvement from baseline in the PtGA of itching decreased in the interrupted group after week 12, but improved between weeks 20 and 24. Notably, mean scores in the PtGA of joint pain of the interrupted group returned to baseline levels (Fig. 2b).

DLQI.  Mean baseline DLQI scores were 12.2 and 12.5 for the continuous and interrupted groups, respectively. At week 12, mean improvements in DLQI scores were 66.9% and 67.7% from baseline, respectively. DLQI improvement was maintained in the continuous group and diminished in the interrupted group after week 12. At week 24, mean improvements in DLQI scores from baseline were 66.6% and 54.0%, respectively.

To assist in the clinical interpretation of these results, we compared these data using a banding method proposed by Hongbo et al. [18], which correlated specific DLQI scores with the effect of psoriasis on patients' lives (DLQI scores 0–1 = no effect; DLQI scores 2–5 = small effect; DLQI scores 6–10 = moderate effect; DLQI scores 11–20 = very large effect; DLQI scores 21–30 = extremely large effect). Based on this method, etanercept therapy provided clinically important improvements in patients' health-related quality of life. Whereas psoriasis had a very large effect on patients' lives in both groups at baseline (mean scores 12.2 and 12.5 for the continuous and interrupted groups, respectively), the disease had a small effect at week 12 (mean scores 3.5 and 3.7, respectively) and week 24 (mean scores 3.6 and 5.2, respectively). Subscale results (symptoms and feelings, daily activities, leisure, work and school, personal relationships, treatment) were generally similar to those observed with the DLQI total score analysis.

SF-36 vitality domain.  Mean baseline SF-36 vitality domain scores were similar in both treatment groups (53.7 for the continuous group and 54.1 for the interrupted group). At week 12, there was a mean percentage improvement from baseline of 76.5% (9.2 points) in the continuous group and 74.0% (8.4 points) in the interrupted group. Improvements in these scores were generally maintained in both groups at week 24 (72.8% [9.6 points] and 70.6% [7.6 points], respectively).

BDI.  The mean BDI scores at baseline were 8.1 and 8.3 in the continuous and interrupted treatment arms, respectively. Mean BDI scores improved from baseline to week 12 by about 29% in both groups (28.8% and 29.3%, respectively). The improvements were generally maintained in the continuous (28.6% [95% CI 23.1–34.1]) and interrupted groups (23.9% [95% CI 18.6–29.1]) at week 24.

EuroQoL-FT.  Mean EuroQoL-FT scores improved from baseline (74.7 for the continuous arm and 74.5 for the interrupted arm) to week 12 in both treatment groups (by 49.0% [8.0 points] and 38.0% [8.2 points] in the continuous and interrupted groups, respectively). These improvements were maintained in both groups at week 24 (52.9% [95% CI 21.4–84.4] and 36.5% [95% CI 20.0–53.0], respectively).

Patient satisfaction survey.  At baseline, more than 60% of patients in both treatment groups indicated that they were not at all satisfied with their previous or current psoriasis treatment, while only 12% of patients in both treatment groups indicated that they were either “very much” or “a lot” satisfied with their psoriasis treatment. After 12 weeks of etanercept treatment, 52.5% and 49.0% of patients in the continuous and interrupted groups, respectively, indicated that they were “very much” satisfied with their psoriasis treatment. Moreover, 77.6% and 77.4% of patients, respectively, rated satisfaction with their psoriasis therapy in the past week at “very much” or “a lot.” These levels of satisfaction were largely maintained through week 24 in the continuous treatment group, although somewhat reduced for some measures in the interrupted group (69.1% vs. 57.6% rated their satisfaction with treatment as “very much” or “a lot” at week 24). Results for the other questions in the satisfaction survey generally followed a similar trend to the result for overall satisfaction with treatment.

Burden of illness.  HRU was significantly reduced from baseline with etanercept treatment in continuous and intermittent treatment groups, respectively, at week 12 in terms of the proportions of patients requiring hospitalization (5.4% vs. 3.6% [P = 0.0106] and 5.7% vs. 3.2% [P = 0.0034]), patients requiring emergency department or urgent care clinic visits (11.2% vs. 8.2% [P = 0.0061] and 12.3% vs. 7.8% [P = 0.0002]), and patients requiring nondermatologist office visits (49.8% vs. 39.5% [P < 0.0001] and 51.8% vs. 37.2% [P < 0.0001]). Improvements were generally maintained at week 24, with no significant difference between the continuous and interrupted groups. Figure 3 shows the mean 24-week reductions in the number of days hospitalized, number of emergency department or urgent care clinic visits (P < 0.005 vs. baseline for both continuous and intermittent groups), and number of nondermatologist office visits (P < 0.005 vs. baseline for both continuous and intermittent groups).

Figure 3.

Mean reduction from baseline in health-care resource utilization (HRU) at week 24. Bootstrap methodology (bootstrap sample: 1000) was used to account for non-normally distributed data. *P < 0.0005 vs. baseline. CI, confidence interval.

The percentages of gainfully employed patients were similar for both groups at baseline (88% for both groups) and at week 12 (no significant change from baseline to week 12: 87% for both groups). Nevertheless, an increase in the percentage of patients working for pay was noted at week 24 relative to baseline for both groups (92.6% and 91.9% for continuous and intermittent groups, respectively; both P ≤ 0.0001). The percentages of patients who had to change job responsibilities because of their psoriasis were improved from baseline to week 12 in both treatment groups and continued to improve at week 24, reaching statistical significance (3.8% at baseline to 1.7% at week 24 [P = 0.001] for the continuous group and 3.5% at baseline to 1.9% at week 24 [P = 0.023] for the interrupted group). The proportions of patients who reported no sick days off from work in the previous month also improved with either continuous or interrupted etanercept therapy: 70.8% vs. 72.6% at baseline, 76.8% vs. 78.3% at week 12, and 74.5% vs. 75.3% at week 24.

Analysis of patients who discontinued and then restarted etanercept therapy.  For the prospective analysis, the interrupted treatment group contained patients who continued etanercept therapy (nonresponders at week 12) and those who discontinued treatment (responders at week 12). To evaluate the PROs in the responders who discontinued therapy but then experienced a relapse and were re-treated, a post hoc analysis was conducted for PtGA of psoriasis and DLQI. A total of 488 patients (53%) who discontinued therapy at week 12 restarted etanercept therapy at week 16 (n = 329) and week 20 (n = 159) (Fig. 1). Mean PtGA of psoriasis and DLQI scores for this subgroup decreased (improved) from baseline (mean PtGA score 4.0; mean DLQI score 12.2) to week 12 (mean PtGA score 1.4; mean DLQI score 3.1) but increased at the re-treatment visit (mean PtGA score 2.7; mean DLQI score 7.0; both P < 0.001 from baseline) (Fig. 4). The mean change from discontinuation at week 12 to the re-treatment visit was less than the minimal clinically important difference of five units for the DLQI [19]. About one-third of patients (167 of 472 patients with evaluable DLQI data) experienced a decline in DLQI score of greater than five units after treatment withdrawal. Furthermore, according to the banding DLQI model proposed by Hongbo et al. [18] that was described earlier, etanercept also had a clinically important effect on these patients. Although psoriasis had a very large effect on patients' lives at baseline (mean score 12.2), the impact of the disease was small at week 12 (mean score 3.1) and moderate at re-treatment (mean score 7.0).

Figure 4.

Values for (a) the patient global assessment (PtGA) of psoriasis and (b) the Dermatology Life Quality Index (DLQI) at week 12 and at re-treatment visit for week-12 responders who discontinued and then restarted etanercept therapy. *< 0.001 vs. baseline. SEM, standard error of the mean.

Discussion

Psoriasis is associated with a significant burden of illness that has a major impact on PROs and HRU [10]. In this study, both continuous and interrupted etanercept treatment improved patient well-being and reduced HRU. Measures of PtGA of psoriasis, joint pain, and itching all were markedly improved and sustained with continuous etanercept treatment, although improvement initially declined in the interrupted group during the discontinuation phase of the study but was gradually regained after re-treatment, with the exception of joint pain. Measures of health-related quality of life, depression, and vitality were improved with treatment in both groups. These results are important because nearly 80% of patients with severe psoriasis indicate that their condition has a negative impact on their lives (including profound emotional, social, and physical impact on quality of life) [8].

Patients were also satisfied with etanercept treatment, although this declined somewhat in the interrupted group during the discontinuation phase of the study. Indeed, the proportion of patients who indicated that they were not at all satisfied with their psoriasis treatment in both groups decreased from more than 60% at baseline to less than 10% at week 12, and the percentage of patients who were “very much” or “a lot” satisfied with their treatment increased from approximately 12% at baseline to nearly 80% at week 12. This is a substantial improvement during the course of therapy and an important outcome as many patients are not satisfied with their psoriasis treatment. A National Psoriasis Foundation (NPF) survey of their patient membership found that 49% of patients were either only somewhat or not at all satisfied with their psoriasis treatments, and 32% of those with severe psoriasis indicated that their treatment was not aggressive enough [8]. By design, the NPF survey captured a much broader patient population than those enrolled into this present study (limited to patients with ≥10% body surface area).

Health-related quality-of-life results in this study are consistent with those from randomized, placebo-controlled clinical trials [11,20–23]. Mean improvement in DLQI scores at week 12 was approximately 70% in a previous study of patients receiving etanercept 50 mg QW and 50 mg BIW. At 24 weeks, mean improvements in DLQI scores were approximately 65% in both dosing arms [20]. In this study, we reported about 67% improvement from baseline in the DLQI at weeks 12 and 24 in the continuous treatment group, and about 70% and 54% improvements in the interrupted group at weeks 12 and 24, respectively. Improvements in the DLQI and other assessments confirmed our initial hypothesis that interrupted treatment would have lower benefits on PRO than continuous etanercept treatment at week 24.

Because not all patients randomized to the interrupted group actually discontinued treatment with etanercept, it was important to determine how patients who withdrew from etanercept at week 12 fared with respect to their PtGA and DLQI responses after discontinuation and before re-treatment. Health-related quality-of-life measures did decrease on discontinuation, but did not reach baseline levels. Therefore, the post hoc hypothesis that health-related quality-of-life measures would decline to baseline levels on abrupt discontinuation of etanercept was not supported. This may be due in part to the longevity of an etanercept effect, and that in addition to relieving the signs and symptoms of psoriasis, etanercept has been shown to improve fatigue and symptoms of depression in a randomized, double-blind, phase 3 study [21]. The results of this post hoc analysis were consistent with the overall results in the interrupted treatment group and suggest that if patients respond well to etanercept treatment but require a holiday period, they may discontinue treatment and continue to benefit from re-treatment at a later time. This clinical observation is supported in the literature for etanercept and other biologics in the treatment of psoriasis [24–26] and for topical corticosteroid treatment of atopic dermatitis [27]. Nevertheless, analysis of DLQI scores indicated that continuous therapy provided more sustained improvements in skin-related health-related quality of life than interrupted therapy, which led to some fluctuations in patients' health-related quality of life.

This is the first study to report the effect of etanercept treatment on work-related outcomes in patients with moderate to severe plaque psoriasis. Patients with well-controlled psoriasis miss fewer days of work, and their psoriasis has less impact on their job responsibilities while in the workplace. Continuous and interrupted therapy with etanercept also significantly reduced the proportions of patients requiring hospitalization, emergency department visits or urgent medical attention, and nondermatologist physician visits. Although costs were not included in the analysis, improvements in these parameters have the potential to reduce direct and indirect medical costs associated with psoriasis. A formal economic analysis is warranted to compare cost effectiveness of the continuous versus intermittent treatment.

One important study limitation is the open-label study design, which introduces the potential for a treatment outcome bias (i.e., patients knew that they were being treated and therefore expected a positive effect). Nevertheless, data from this study are in line with observations from placebo-controlled trials of etanercept in psoriasis. Another consideration is that the changes in work-related outcomes and HRU may have been due in part to the increased monitoring and clinician oversight that accompanies a clinical trial. These results will need to be verified in further trials. And finally, patients in the interrupted treatment group may have required additional time to experience the full effects of etanercept after discontinuation and re-treatment.

Continuous treatment with etanercept 50 mg BIW for 12 weeks followed by 50 mg QW for 12 weeks provided sustained and clinically important improvements in PROs and reductions in HRU. Although there were reductions in some outcome measures after treatment discontinuation at week 12 in the interrupted group, most changes did not revert to baseline levels, and improvements similar to week 12 levels were observed on re-treatment. In conclusion, continuous treatment of etanercept does provide greater improvements in PROs than does interrupted treatment. Should the need arise to interrupt etanercept therapy, one can expect a slow reduction in PRO and HRU benefits and reclaimed benefit after reinstituting etanercept treatment.

The authors would like to thank Rick Davis, MS, RPh, on behalf of Amgen Inc., for writing assistance.

Source of financial support: Research was funded by Immunex Corporation, a wholly owned subsidiary of Amgen Inc., and by Wyeth Pharmaceuticals.

Ancillary