To evaluate the impact of adalimumab on health-related quality of life (HRQOL) in patients with active ankylosing spondylitis (AS).
To evaluate the impact of adalimumab on health-related quality of life (HRQOL) in patients with active ankylosing spondylitis (AS).
Patients ≥18 years enrolled in the Adalimumab Trial Evaluating Long-Term Efficacy and Safety in Ankylosing Spondylitis, a randomized controlled study, were randomly assigned to receive either adalimumab 40 mg subcutaneously or placebo every other week for 24 weeks. ASsessment of Ankylosing Spondylitis (ASAS) International Working Group criteria were used to evaluate clinical efficacy. HRQOL outcomes were assessed using the Short Form 36 (SF-36) Health Survey and Ankylosing Spondylitis Quality of Life (ASQoL) Questionnaire.
A total of 315 patients enrolled (208 in the adalimumab group and 107 in the placebo group). Patients in the adalimumab group showed significant improvements in SF-36 Physical Component Summary (PCS) and ASQoL scores versus placebo at weeks 12 and 24 (P < 0.001). The observed differences between adalimumab and placebo patients exceeded the a priori minimum important difference (MID) at the group level, and significantly more adalimumab-treated patients achieved improvements greater than the MID on the patient level. These data suggest the HRQOL improvements were clinically meaningful. No differences were observed in SF-36 Mental Component Summary (MCS) scores. Significant differences favoring adalimumab were observed for SF-36 domains physical function, bodily pain, role-physical, general health, vitality, social function, and role–emotional. There was significant association between HRQOL improvements (measured by SF-36 PCS and MCS, and ASQoL scores) and ASAS clinical responses (P < 0.001).
Adalimumab significantly improved physical health status and overall HRQOL through 24 weeks in patients with active AS.
Ankylosing spondylitis (AS) is a chronic, progressive, inflammatory disease primarily affecting the axial skeleton, peripheral joints, and entheses. The prevalence of AS is reported to be between 0.1% and 0.8%, with more males affected than females (1). AS is most often diagnosed between the ages of 20 and 40 years. As the disease progresses, patients with AS experience pain, joint stiffness, and the eventual loss of spinal mobility. These clinical symptoms and subsequent disease progression result in functional limitations and impairment in health-related quality of life (HRQOL) (2–5) and work productivity (6, 7).
Until the last few years, there were only a limited number of treatments available for AS, including nonsteroidal antiinflammatory drugs (NSAIDs) and physiotherapy, both of which provide short-term symptomatic relief. Traditional disease-modifying antirheumatic drugs (DMARDs), such as methotrexate and sulfasalazine, have demonstrated limited effectiveness. Recent studies have shown that AS is an immunologically mediated disease in which tumor necrosis factor (TNF) is present in increased concentrations in joints (8). Joint recommendations of the ASsessment in Ankylosing Spondylitis International Working Group (ASAS) and the European League Against Rheumatism state that “anti-TNF treatment should be given to patients with persistently high disease activity despite conventional treatments according to ASAS recommendations. There is no evidence to support the obligatory use of DMARDs before, or concomitant with, anti-TNF treatment in patients with axial disease” (9). These recommendations are a result of the large improvements in pain and function that have been achieved in clinical trials of the 3 TNF antagonists approved for the treatment of AS (9).
Adalimumab is a fully human, anti-TNF monoclonal antibody with high affinity and specificity for human TNF, which has demonstrated efficacy in patients with rheumatoid arthritis (10–12) and psoriatic arthritis (13). The recently completed Adalimumab Trial Evaluating Long-Term Efficacy and Safety in Ankylosing Spondylitis (ATLAS) study also found that adalimumab was well tolerated and efficacious in treating patients with AS (14). Moreover, adalimumab received regulatory approvals for the treatment of AS in the US and Europe in June and July 2006, respectively.
Patient-reported outcomes provide important assessments of functioning and well-being from the patients' perspective, which complement and expand on traditional clinical outcomes in AS (15–17). AS impacts several HRQOL domains (5), including pain, physical function, fatigue, and psychological well-being (2–4, 17–20). The objective of this study was to assess the effects of adalimumab on HRQOL in patients with active AS in the ATLAS study.
Study patients were from the ATLAS trial, a multicenter, double-blind, phase III, randomized, placebo-controlled study designed to demonstrate the safety and efficacy of adalimumab in patients with active AS. Details about the study population selection, inclusion/exclusion criteria, and other basic study information are published separately (14). Study patients were recruited from 43 sites (21 in the US and 22 in Europe). Adults 18 years and older were eligible for study participation if they were diagnosed with AS according to the modified New York criteria (21) and had an inadequate response to or intolerance of ≥1 NSAID. Patients for whom ≥1 DMARD had failed were also allowed to participate.
Patients were randomized in a 2:1 ratio to receive either adalimumab 40 mg or matching placebo subcutaneously every other week for 24 weeks. Patients who did not achieve a 20% response according to the ASAS criteria for improvement (ASAS20) (22) at weeks 12, 16, or 20 were eligible to receive early-escape, open-label treatment with adalimumab 40 mg every other week, and were still followed per the clinical trial protocol (14).
The ASAS20, ASAS40, and ASAS5/6 response criteria (22, 23) were used to evaluate clinical efficacy (14). The ASAS20 response criteria consist of at least 20% and absolute improvement of at least 1 unit versus baseline in at least 3 of the following 4 domains: patient's global assessment, pain, function (represented as the Bath Ankylosing Spondylitis Functional Index [BASFI] score) (24), and inflammation (represented as the mean of the 2 morning stiffness–related Bath Ankylosing Spondylitis Disease Activity Index [BASDAI] scores) (25), with no deterioration (defined as a worsening of ≥20% or an absolute increase of at least 1 unit) in the remaining domain. The ASAS40 response represents improvement of at least 40% and absolute improvement of at least 2 units versus baseline in at least 3 of the 4 domains of the ASAS20 criteria, with no deterioration in the remaining domain. The ASAS5/6 criteria used C-reactive protein level and spinal mobility (Bath Ankylosing Spondylitis Metrology Index [BASMI] score) (26) in addition to the 4 domains of the ASAS20. The ASAS5/6 criteria are fulfilled if there are ≥20% improvements in 5 of 6 domains. In ATLAS, ASAS response was used in planned exploratory analyses of the relationship between the HRQOL outcomes and clinical responses. Patients were categorized into mutually exclusive groups corresponding to ASAS20, ASAS40, and ASAS5/6 response criteria, based on data from the 12-week visit.
HRQOL measures represent multidomain constructs that include physical, psychological, social functioning, and well-being dimensions, but can also include 1 or more single-dimensional assessments of disease-relevant symptom outcomes (27, 28). In ATLAS, HRQOL was measured by the Short Form 36 Health Survey (SF-36) at baseline and weeks 12 and 24, and by the Ankylosing Spondylitis Quality of Life (ASQoL) questionnaire at baseline and weeks 2, 12, and 24.
The SF-36 is a generic health status instrument developed for application in primary care and chronic disease populations (29). The SF-36 version 1, with a 4-week recall period, was used in this study. The SF-36 has 2 summary scores, the Physical Component Summary (PCS) and Mental Component Summary (MCS) scores, and 8 domain scores: physical function, bodily pain, role limitations–physical, general health, vitality, social function, role limitations–emotional, and mental health. Domain scores range from 0 to 100, with greater scores reflecting better health status. The summary scores are constructed as T scores with a mean ± SD of 50 ± 10. Greater scores again indicate better health status. The SF-36 summary and domain scores have excellent reliability and good construct validity across the general US population as well as chronic disease populations (29, 30), including patients with AS (2, 16, 18, 31, 32). Norms for the general US population are also available (27). Published research indicates that a difference of 2.5–5 points in SF-36 PCS or MCS scores and 5–10 points in SF-36 domain scores is considered clinically important (33, 34). In this analysis, a change of ≥3 points was specified a priori as the minimum important difference (MID) criteria for SF-36 PCS.
Developed from a needs-based model (35), the ASQoL is a disease-specific instrument designed to measure HRQOL in patients with AS. Patients answer yes/no to 18 items assessing the current impact of AS on their quality of life status. The ASQoL has a total score ranging from 0 to 18, with lower scores representing better AS-specific quality of life. The ASQoL has demonstrated good reliability and construct validity across several different AS studies (36). Published research has indicated that differences of 1–2 points in ASQoL scores are clinically significant (17, 35–39). In this analysis, a change of ≥1.8 points, which represents 10% of the possible score range, was specified a priori as the MID criteria for ASQoL score.
The primary HRQOL end points were identified a priori as the baseline to 12-week and 24-week mean changes in SF-36 PCS and ASQoL scores. Secondary end points included the SF-36 MCS and SF-36 domain scores over 24 weeks. These analyses were conducted for a modified intent-to-treat population defined as all randomized patients who received at least 1 dose of study treatment and completed baseline HRQOL assessments and at least 1 followup HRQOL assessment within 24 weeks of study entry. Last observation carried forward (LOCF) procedures were used to handle HRQOL scores after patients discontinued from the study or went into early-escape therapy.
Baseline demographics, clinical characteristics, and HRQOL measures were summarized by treatment group using descriptive statistics. One-way analysis of variance and Fisher's exact tests were performed to examine whether there were significant differences between treatment groups at baseline.
Analysis of covariance (ANCOVA) was performed to compare mean baseline HRQOL scores with week 12 and week 24 HRQOL change scores between the treatment groups. The ANCOVA models included terms for treatment group and baseline assessments. The analysis of primary HRQOL end points was completed based on a hierarchical, closed-test procedure to handle multiplicity in the HRQOL end points. No adjustments in alpha level were needed at each stage of statistical testing. All statistical tests were 2 sided and were performed at an alpha level ≤0.05. The test procedure first evaluated treatment differences via mean PCS scores from baseline to 24 weeks between the adalimumab and placebo groups. Then, if the first test was statistically significant, the procedure evaluated treatment differences via mean changes in ASQoL scores from baseline to 24 weeks.
A repeated-measures ANCOVA model was used to compare mean SF-36 (including PCS, MCS, and domain scores) and ASQoL scores collected at the primary visits (i.e., baseline and weeks 12 and 24) between treatment groups (36). The ANCOVA models included terms for treatment group, week, treatment-by-week interaction, and relevant baseline score (as covariates). Statistical analyses of the SF-36 MCS and domain scores were considered supportive and exploratory.
An exploratory data analysis was performed to evaluate the relationship between changes in the HRQOL measures and the ASAS responses. ANCOVA models were used to estimate least squares mean baseline to 12-week change scores for the SF-36 PCS, SF-36 MCS, and ASQoL scores. The ANCOVA models included factors for ASAS response groups (i.e., <ASAS20, ≥ASAS20 but <ASAS40, and ≥ASAS40), age, sex, and the relevant baseline scores. A separate ANCOVA examined differences between the groups meeting and not meeting ASAS5/6 response criteria. The exploratory analysis focused on changes from baseline to week 12 rather than baseline to week 24, because a number of patients transitioned to early-escape treatment before their week 24 visits.
The primary data analyses were conducted by the authors at the United BioSource Corporation (UBC; authors DR and AMR), independent of the pharmaceutical company that sponsored the study. Several authors were responsible for designing the study, which included selection of the quality of life instruments used (JCD, DMFvdH, RLW, HK). All authors took active roles in interpreting the data. The UBC authors also had the lead role in drafting the manuscript. All authors critically reviewed the manuscript, made substantial contributions to its development, and approved/provided consent that the final version be submitted.
A total of 315 patients with active AS participated in this study: 208 in the adalimumab group and 107 in the placebo group. Patient demographics and clinical characteristics are summarized in Table 1. Most patients were white (95.6%) and male (74.9%), and the mean age was 42.2 years. Patients had a mean disease duration of 10.6 years. Baseline demographics and clinical characteristics were similar between treatment groups.
|Variables||Placebo (n = 107)||Adalimumab (n = 208)||P†|
|Age, years||43.4 ± 11||41.7 ± 12||NS|
|Male sex, no. (%)||79 (74)||157 (76)||NS|
|White, no. (%)||99 (93)||202 (97)||0.05|
|Disease duration, years||10.0 ± 8||11.3 ± 10||NS|
|HLA–B27 positive||85 ± 79||163 ± 78||NS|
|BASDAI score (0–10-cm VAS)||6.3 ± 2||6.2 ± 2||NS|
|BASFI score (0–10-cm VAS)||5.6 ± 2||5.2 ± 2||NS|
Baseline scores for the SF-36 PCS, MCS, and domain scores, as well as ASQoL scores are presented in Table 2. Overall, there were no statistically significant differences in baseline HRQOL scores between the treatment groups. Baseline SF-36 PCS scores (placebo: 31.8; adalimumab: 32.9) were almost 20 points lower than the US general population norm of 50, indicating substantial impairment of physical health status. Baseline SF-36 MCS scores (placebo: 44.4; adalimumab: 43.4) were close to the population norm. Significantly lower scores relative to the US population norms were also found in all SF-36 domains, with the exception of mental health.
|Measure||Placebo (n = 107)||Adalimumab (n = 208)||P†|
|SF-36 PCS||31.8 ± 8||32.9 ± 8||NS|
|SF-36 MCS||44.4 ± 12||43.4 ± 12||NS|
|SF-36 domain scales|
|Physical function||45.5 ± 23||47.9 ± 22||NS|
|Role–physical||19.9 ± 31||20.5 ± 30||NS|
|Social function||53.4 ± 26||57.2 ± 25||NS|
|General health||41.0 ± 19||43.4 ± 20||NS|
|Bodily pain||29.8 ± 15||31.7 ± 17||NS|
|Vitality||34.0 ± 17||32.6 ± 18||NS|
|Role–emotional||56.8 ± 43||53.2 ± 43||NS|
|Mental health||62.5 ± 19||61.3 ± 19||NS|
|ASQoL||10.6 ± 4||10.2 ± 4||NS|
Almost all adalimumab-treated patients (98.1%) and placebo-treated patients (96.3% completed the 12-week double-blind study period) (14). Of 315 patients, 296 (94%) remained in the study by week 24, but only 29 of 107 patients randomized to placebo completed week 24 while continuing to receive placebo. A total of 155 patients (49.2%) received early-escape, open-label therapy (21 of 208 adalimumab-treated patients, 74 of 107 placebo-treated patients) (14).
Descriptive statistics of the mean change scores for the SF-36 PCS and ASQoL over time are presented in Table 3. Overall, adalimumab-treated patients demonstrated statistically significant improvements in SF-36 PCS scores (P < 0.001) and ASQoL scores (P < 0.001) compared with placebo-treated patients at week 12, and these improvements were sustained through 24 weeks. In addition, HRQOL improvements, as measured by the ASQoL, occurred as early as 2 weeks after start of adalimumab treatment (−1.9 for adalimumab versus –0.1 for placebo; P < 0.001).
|Measure||Baseline to week 12||Baseline to week 24|
|Placebo (n = 107)||Adalimumab (n = 208)||P†||Placebo (n = 107)||Adalimumab (n = 208)||P†|
|SF-36 PCS||1.6 ± 0.8||6.9 ± 0.6||< 0.001||1.9 ± 0.9||7.4 ± 0.6||< 0.001|
|SF-36 MCS||2.4 ± 1.0||2.7 ± 0.7||NS||2.4 ± 1.0||3.6 ± 0.7||NS|
|ASQoL||−1.0 ± 0.4||−3.2 ± 0.3||< 0.001||−1.1 ± 0.4||−3.6 ± 0.7||< 0.001|
HRQOL improvements with adalimumab were considered clinically meaningful, because the mean change from baseline in SF-36 PCS scores at both weeks 12 and 24 exceeded the a priori MID of 3 points for the adalimumab group but not for the placebo group. Moreover, the between-group differences (adalimumab versus placebo) in SF-36 PCS scores exceeded the MID criteria at both weeks 12 and 24. Significantly more adalimumab-treated patients achieved the MID versus placebo-treated patients at week 12 (65.0% versus 37.6%; P < 0.001) and week 24 (67.3% versus 39.6%; P < 0.001). Similarly, clinically meaningful improvements were observed in ASQoL scores. The mean change from baseline scores in adalimumab-treated patients, as well as the between-group differences (adalimumab versus placebo), exceeded the a priori MID of 1.8 points. In addition, 59.6% of adalimumab patients versus 42.1% of placebo patients achieved the MID criteria at week 12 (P < 0.01), and 65.4% versus 42.1% (P < 0.001) did so at week 24.
Repeated-measure ANCOVAs were performed to compare mean SF-36 PCS and ASQoL scores between treatment groups. Overall, there were significant treatment-by-week interactions for both SF-36 PCS and ASQoL scores (P < 0.001) and significant main effects of treatment (P < 0.001).
The adalimumab treatment arm showed statistically significant improvements in mean SF-36 domain scores, including physical function (P < 0.001), role–physical (P < 0.001), bodily pain (P < 0.001), vitality (P < 0.01), general health (P < 0.001), and role–emotional scores (P < 0.05) at week 12. The exceptions were the social functioning and mental health domain scores (Figure 1). This improvement was maintained through week 24. In addition, the social functioning domain scores reached statistical significance at week 24 (P < 0.01) (data not shown). In all cases, the significant differences favored the adalimumab group, and each was considered clinically meaningful based on the 5–10-point MID criteria reported by Kosinski et al (34). No significant differences were observed between the treatment groups for SF-36 MCS scores at weeks 12 and 24 (Table 3).
Based on the repeated-measures ANCOVA assessments, there were statistically significant treatment-by-week interactions (P < 0.05) for all SF-36 domain scores except social function (P > 0.05 or not significant [NS]) and mental health (NS). There were significant main effects of treatment (P < 0.001) for all SF-36 domain scores except vitality (NS), role–emotional (NS), and mental health (NS).
Exploratory analyses were performed to examine the relationship of baseline to week 12 changes in SF-36 PCS, SF-36 MCS, and ASQoL scores with ASAS response criteria, after adjusting for age and sex. Results of the analyses using the ASAS20 and ASAS40 improvement criteria are presented in Table 4. Overall, there were statistically significant differences in SF-36 PCS, SF-36 MCS, and ASQoL scores between the nonresponder, ASAS20 responder, and ASAS40 responder groups (P < 0.001). Changes in these 3 end points were smallest for patients in the nonresponder group and were greatest for those in the ASAS40 responder group. Changes in the ASAS40 group were more than twice as large as changes observed in the ASAS20 group. Results using the ASAS5/6 response criteria are presented in Table 5. Improvement in SF-36 PCS, SF-36 MCS, and ASQoL scores in the ASAS5/6 responder group was significantly greater compared with the nonresponder group (P < 0.001).
|Measure||ASAS20 and ASAS40 improvement status||Overall P value¶|
|Nonresponder (<ASAS20)†||≥ASAS20 but <ASAS40‡||≥ASAS40§|
|SF-36 PCS||0.7 ± 6.4||5.2 ± 7.1||12.9 ± 7.1||< 0.001|
|SF-36 MCS||1.0 ± 10.9||1.9 ± 9.7||5.6 ± 11.1||< 0.001|
|ASQoL||−0.4 ± 2.7||−2.3 ± 2.3||−6.0 ± 4.0||< 0.001|
|Measure||ASAS5/6 improvement status||Overall P value§|
|ASAS5/6 nonresponder†||ASAS5/6 responder‡|
|SF-36 PCS||1.8 ± 7.4||11.0 ± 7.7||< 0.001|
|SF-36 MCS||0.2 ± 10.5||6.1 ± 10.8||< 0.001|
|ASQoL||−0.7 ± 2.8||−5.2 ± 4.1||< 0.001|
Adalimumab treatment significantly reduced the signs and symptoms of active AS and established a sustained clinical response in patients who had an inadequate response or intolerance to NSAID therapy (14). This study also demonstrates that these clinical benefits favorably impact patients' HRQOL.
At study entry, patients with AS had significant HRQOL impairment compared with the US general population. Disease impact was most significant in physical health status, as demonstrated by baseline SF-36 PCS scores. Twelve-week improvements in overall physical function (SF-36 PCS) and quality of life, as measured by the ASQoL, were significantly greater for adalimumab- versus placebo-treated patients. These improvements were maintained throughout the 24-week double-blind study period, and were found to be clinically meaningful at both the group and individual patient levels, based on the a priori MID for both SF-36 PCS and ASQoL scores. These results suggest a broad and clinically important impact of adalimumab on HRQOL for patients with AS.
There was a significant association between the improvement in HRQOL measures and clinical responses, as measured by the ASAS response criteria. The largest effects were observed for the associations of physical health status (measured by SF-36 PCS) and overall quality of life (measured by ASQoL) with clinical responses, while mental health status (measured by SF-36 MCS) was also found to be related to clinical responses. These findings provide additional support for the clinical responsiveness of these HRQOL end points in patients with AS, and indicate that adalimumab treatment resulted in significant and clinically meaningful HRQOL improvements in patients with AS.
The SF-36 MCS and the mental health domain scores were not significantly different between the treatment groups over the 24-week study period. This is not surprising because the impairment in mental health status of patients with AS (measured by the baseline SF-36 MCS) was less severe than the impairment observed in physical health status and was close to the mental health status reported in the US general population. This result is similar to a recent publication demonstrating that SF-36 MCS scores did not differ significantly between patients treated with either infliximab or placebo (16).
Previous reports of HRQOL results in patients with AS treated with TNF antagonists have been limited in scope. Published results from 2 studies (a 16-week randomized controlled trial with 40 patients and a 48-week open-label extension with 129 patients) reported improvements in AS with etanercept therapy in SF-36 domains (18). In addition, HRQOL results from infliximab randomized controlled trials have been limited to SF-36 component summary scores (14, 40). Although trials cannot be directly compared, the magnitude of improvements in SF-36 measures between all 3 TNF antagonists has been similar. However, in addition to demonstrating statistically significant differences in HRQOL measures between adalimumab and placebo, the findings in this analysis also demonstrate the clinical meaningfulness of these data at both the patient and group levels. In contrast to the reports on etanercept and infliximab therapy, the strong correlations of HRQOL improvements with clinical response such as ASAS20, ASAS40, and ASAS5/6 are presented for adalimumab therapy. Moreover, the present analysis included the first reported results of improvements in overall quality of life for patients with AS through the ASQoL, a validated, AS-specific HRQOL measure (35).
In reviewing the results of this analysis, readers might consider that the clinical differences (e.g., ASAS20) between adalimumab and placebo at week 24 are likely to have been underestimated, primarily because the intent-to-treat analysis mandated that all patients be designated as nonresponders after they entered into early-escape, open-label therapy. Moreover, the 94% retention rate (195 of 208 patients) at week 24 of patients receiving adalimumab further suggests that clinical efficacy was underestimated.
A few limitations associated with the HRQOL assessment in this study should be considered when interpreting these findings. First, a substantial percentage (49.2%) of randomly assigned patients entered into early-escape therapy with open-label adalimumab after 12 weeks. Although LOCF analysis (involving the last available HRQOL assessment while receiving double-blind therapy) was used, the large numbers of patients discontinuing randomized study treatment may compromise the interpretation of the HRQOL results after the 12-week visit. Second, the HRQOL end points were based on patient reports, and it is unknown whether expectations for improvements in clinical and functional outcomes influenced these results. However, a consistent and significant relationship between degree of clinical response and changes in SF-36 PCS and ASQoL scores was observed in this study.
In summary, the results of this study provide evidence supporting the ability of adalimumab to provide patients with statistically significant and clinically meaningful improvements in measures of physical function, as well as AS-specific quality of life outcomes. In addition, adalimumab-treated patients with AS reported consistent and significant improvements across a range of physical function and HRQOL outcomes. Therefore, adalimumab is efficacious in improving physical function and AS-specific quality of life in patients with AS. Combined with the previously reported findings on tolerability and clinical efficacy with adalimumab (14), these results suggest that adalimumab improves physical function and overall well-being, and may be a comprehensively effective therapy for patients with AS.
Dr. Davis had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Davis, Revicki, van der Heijde, Wong, Kupper.
Acquisition of data. Davis, van der Heijde, Wong, Kupper.
Analysis and interpretation of data. Davis, Revicki, van der Heijde, Rentz, Wong, Kupper, Luo.
Manuscript preparation. Davis, Revicki, van der Heijde, Rentz, Wong, Luo, Michael A. Nissen (nonauthor; Abbott Laboratories).
Statistical analysis. Rentz, Luo.
The ATLAS trial and this analysis were fully funded by Abbott Laboratories. The design of the study, including the selection of quality of life instruments, was determined by an international steering committee (including the authors JD and DvdH) as well as a few Abbott employees (RW, ML, and HK). The Abbott authors, as well as an Abbott medical writer assisted in the development and revision of the manuscript. The Abbott authors, in conjunction with all other authors, critically reviewed the manuscript and agreed to submit it for publication.
The authors thank Eric D. Bauer, BS, Shyanne M. Douma, BS, and Hans H. Kissel, PhD, as well as all of the ATLAS study investigators, and the study site coordinators, for assistance in conducting the study.