To evaluate the effect of golimumab on sleep disturbance in patients with active ankylosing spondylitis (AS).
To evaluate the effect of golimumab on sleep disturbance in patients with active ankylosing spondylitis (AS).
Golimumab was studied in a multicenter, randomized, placebo-controlled study (GO-RAISE). At baseline, 356 patients were randomly assigned in a 1.8:1.8:1 ratio to subcutaneous golimumab 50 mg, 100 mg, or placebo every 4 weeks. Sleep disturbance was assessed using the Jenkins Sleep Evaluation Questionnaire (JSEQ), which was administered at baseline, week 14, and week 24. Treatment effect was evaluated using analysis of variance on the van der Waerden normal scores.
Median JSEQ scores at baseline were 9.0 in the placebo group, 10.0 in the 50-mg group, and 11.0 in the 100-mg group, indicating moderate to severe sleep disturbance. Patients who received golimumab showed significantly greater median improvement from baseline in JSEQ scores compared with placebo at week 14 (−3.0 versus 0.0; P < 0.001) and week 24 (−3.0 versus −1.0; P < 0.001). Changes from baseline in JSEQ scores significantly correlated with changes from baseline in Short Form 36 summary scores, Bath AS Functional Index scores, total back pain, night back pain, and Bath AS Disease Activity Index scores. Multiple regression analyses indicated that improvement in the night back pain score was the most consistent predictor of change in JSEQ score or reduction in sleep disturbance.
Patients with active AS showed significant sleep disturbance at baseline due to underlying pain associated with AS. Treatment with subcutaneous golimumab every 4 weeks significantly reduced sleep disturbance and improved health-related quality of life.
Ankylosing spondylitis (AS) is characterized by inflammation of the sacroiliac joints, axial skeleton, and entheses. Sleep disturbance due to axial pain and stiffness in the latter half of the night is an important characteristic of the inflammatory back pain experienced by patients with AS (1). Patients with AS often report getting out of bed and walking around to reduce pain and stiffness, which can lead to daytime fatigue (1). Although fatigue was once an ignored aspect of the disease (2), recent studies have specifically evaluated fatigue (3) and the effect of AS treatment on fatigue (4). In addition, patients with AS have indicated that although their fatigue and sleep disturbance are related, they are separate entities and should be evaluated as such (5). However, few studies have directly evaluated sleep disturbance and its relationship to fatigue in patients with AS.
A Swedish study (6) used the Uppsala Sleep Inventory questionnaire (7), an 87-question instrument that evaluates sleep and sleep-related issues, to evaluate sleep disturbance in outpatients with AS. In a Turkish study by Karapolat and colleagues (8), the effects of various exercise regimens on disease activity, quality of life, and other outcomes, including sleep, were evaluated in patients with AS. Karapolat and colleagues used the Nottingham Health Profile (9) to evaluate sleep. In a more recent Turkish study of the characteristics of fatigue in patients with AS (10), sleep disturbance was evaluated using the Pittsburgh Sleep Quality Index (11). To our knowledge, no studies have evaluated the effect of pharmacologic intervention on sleep disturbance in patients with AS.
The GO-RAISE study was a randomized, placebo-controlled study that was designed to evaluate the efficacy of golimumab in patients with AS. The effect of golimumab on sleep disturbance was evaluated as one of the secondary outcomes in the GO-RAISE study. Here we report the results of this analysis.
The details of the GO-RAISE study, including eligibility criteria and randomization scheme, have been published previously (12). Briefly, patients with AS for ≥3 months, a Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) score of ≥4, a total back pain score of ≥4, and an inadequate response to current or previous nonsteroidal antiinflammatory drugs or disease-modifying antirheumatic drugs were eligible for the study.
Patients were randomly assigned in a 1:1.8:1.8 ratio to receive subcutaneous placebo (n = 78) or golimumab 50 mg (n = 138) or 100 mg (n = 140) every 4 weeks. Patients, investigators, and study personnel were blinded to treatment assignment. At week 16, patients who had <20% improvement from baseline in both total back pain and morning stiffness entered an early escape arm in which patients in the placebo group received golimumab 50 mg and patients in the golimumab 50-mg group had their dose increased to 100 mg. Patients who were originally assigned to 100 mg may have met the criteria for early escape but did not receive dose adjustments.
Sleep disturbance was evaluated using the Jenkins Sleep Evaluation Questionnaire (JSEQ) (13) (Table 1). The JSEQ assesses the frequency of sleep disturbance in 4 categories: 1) trouble falling asleep, 2) waking up several times during the night, 3) having trouble staying asleep (including waking up far too early), and 4) waking up after the usual amount of sleep feeling tired and worn out. Patients reported the numbers of days they experienced each of these problems in the past month. The numbers of days were categorized and assigned a score from 0 to 5 according to the official scale (Table 1). Thus, the total JSEQ score ranges from 0 to 20, with higher scores indicating greater sleep disturbance.
|How often in the past month did you:|
|1. Have trouble falling asleep?|
|2. Wake up several times per night?|
|3. Have trouble staying asleep (including waking far too early)?|
|4. Wake up after your usual amount of sleep feeling tired and worn out?|
The primary study end point was the proportion of patients who achieved ≥20% improvement in the Assessment of SpondyloArthritis international Society criteria (ASAS20) and has been previously reported (12). Sample size calculations were based on the primary end point analyses; secondary outcomes such as JSEQ scores were not considered in sample size calculations. BASDAI (14) was also used to evaluate disease activity. Physical function was evaluated using the Bath Ankylosing Spondylitis Functional Index (BASFI) (15). Patients scored the amount of night back pain and total back pain they had during the previous week on a 10-cm visual analog scale. Health-related quality of life (HRQOL) was assessed using the Short Form 36 (SF-36) health survey (16, 17).
Changes from baseline to weeks 14 and 24 for continuous variables, including the JSEQ scores, were compared between the treatment groups using an analysis of variance on the van der Waerden normal scores at a nominal level of significance of 0.05. Analyses were conducted on observed data without imputation for missing values. For patients in the golimumab 50-mg group or the placebo group who entered early escape, the last observation before their dose adjustment at week 16 was used for the week 24 analysis. Patients in the 100-mg group were not eligible to receive a dose adjustment according to the protocol; therefore, observed values at week 24 were used for these patients.
Associations between changes from baseline in JSEQ scores and SF-36 physical component summary (PCS) and mental component summary (MCS) scores, BASFI score, night back pain score, total back pain score, BASDAI score, and inflammation score (average of the last 2 questions of the BASDAI regarding morning stiffness) were evaluated using Spearman's rank correlations and multiple regression analysis. The chi-square test was used for comparing the proportion of patients who achieved a decrease of ≥2 in the score for each individual JSEQ question.
The unidimensionality of the JSEQ was assessed by Cronbach's alpha coefficient. Aspects of truth according to the Outcome Measures in Rheumatology Clinical Trials filter (18) were assessed by evaluating correlations with clinical and patient-reported outcomes and responder analysis.
The primary results of the GO-RAISE study have been previously published (12). Relevant baseline assessments for the present analysis are summarized in Table 2. The treatment groups were generally well balanced for each of the major clinical assessments. However, the median time since diagnosis of AS was generally longer in patients in the placebo group (7 years) compared with those in the golimumab groups (5 years). The mean baseline JSEQ score was 9.9 in the placebo group, 10.3 in the 50-mg golimumab group, and 11.1 in the 100-mg golimumab group.
|Placebo (n = 78)||Golimumab|
|50 mg (n = 138)||100 mg (n = 140)|
|Men, no. (%)||55 (70.5)||102 (73.9)||98 (70.0)|
|Years since diagnosis of AS|
|Mean ± SD||11 ± 10||8 ± 8||8 ± 8|
|Median (IQR)||7 (3–19)||5 (2–12)||5 (2–13)|
|JSEQ score (0–20)|
|Mean ± SD||9.9 ± 4.7||10.3 ± 4.4||11.1 ± 4.8|
|Median (IQR)||9.0 (6.0–14.0)||10.0 (7.0–14.0)||11.0 (8.0–15.0)|
|Inflammation (morning stiffness; 0–10)|
|Mean ± SD||6.8 ± 2.0||6.7 ± 1.9||7.4 ± 1.9|
|Median (IQR)||7.1 (5.5–8.3)||7.1 (5.4–8.1)||7.6 (6.1–9.0)|
|Duration of morning stiffness, minutes|
|Mean ± SD||76 ± 31||76 ± 31||86 ± 31|
|Median (IQR)||77 (46–104)||77 (53–100)||90 (60–118)|
|BASFI score (0–10)|
|Mean ± SD||5.1 ± 2.3||5.0 ± 2.4||5.2 ± 2.6|
|Median (IQR)||4.9 (3.5–6.8)||5.0 (3.2–6.7)||5.4 (3.4–7.3)|
|Night back pain score (0–10)|
|Mean ± SD||6.9 ± 2.3||6.5 ± 2.2||7.2 ± 2.1|
|Median (IQR)||7.4 (6.0–8.6)||7.1 (5.2–8.1)||7.6 (6.5–8.8)|
|Total back pain score (0–10)|
|Mean ± SD||7.5 ± 1.6||7.1 ± 1.5||7.6 ± 1.6|
|Median (IQR)||7.6 (6.6–8.8)||7.5 (5.7–8.2)||7.9 (6.5–8.8)|
|SF-36 PCS score (0–100)|
|Mean ± SD||29.2 ± 7.1||29.3 ± 8.3||30.7 ± 8.1|
|Median (IQR)||28.3 (23.8–34.1)||29.7 (22.5–35.3)||29.8 (25.2–35.5)|
|SF-36 MCS score (0–100)|
|Mean ± SD||45.7 ± 10.7||45.1 ± 10.5||42.9 ± 11.5|
|Median (IQR)||46.2 (37.1–54.8)||46.5 (36.8–54.1)||43.1 (33.5–53.5)|
The JSEQ is a unidimensional instrument for assessing sleep in patients with AS. The Cronbach's alpha coefficient was 0.72. The correlations assessing truth were in the hypothesized directions, showing that reductions in sleep disturbance correlated with improvements in clinical and patient-reported outcomes (Table 3). ASAS20 responders had greater median improvement in JSEQ scores than ASAS20 nonresponders at weeks 14 (4.0 versus 1.0; P < 0.001) and 24 (4.0 versus 1.0; P < 0.001).
|Measures||Week 14||Week 24|
|PCS score||−0.35||< 0.001||−0.27||< 0.001|
|MCS score||−0.23||< 0.001||−0.21||< 0.001|
|Vitality score||−0.36||< 0.001||−0.35||< 0.001|
|BASFI score||0.37||< 0.001||0.37||< 0.001|
|Night back pain score||0.36||< 0.001||0.40||< 0.001|
|Total back pain score||0.33||< 0.001||0.36||< 0.001|
|BASDAI score||0.39||< 0.001||0.38||< 0.001|
|Fatigue score||0.41||< 0.001||0.38||< 0.001|
|Inflammation (morning stiffness) score||0.33||< 0.001||0.38||< 0.001|
Patients in the golimumab 50- and 100-mg groups had a significantly greater reduction from baseline in JSEQ scores at week 14 compared with the placebo group (P < 0.001) (Figure 1A). This improvement was sustained at week 24 and was similar in each of the individual golimumab dose groups.
As reported previously, 22% of patients in the placebo group, 59% of patients in the 50-mg group, and 60% of patients in the 100-mg group achieved the primary end point of the study, an ASAS20 response at week 14 (12). Even among the subset of patients who achieved an ASAS20 response, the golimumab groups showed significantly greater reductions in JSEQ scores than the placebo group (Figure 1B). There was no statistically significant difference between the treatment groups among the subset of ASAS20 nonresponders (Figure 1C).
Golimumab-treated patients showed improvement in each of the individual questions of the JSEQ. The proportion of patients with an improvement in ≥2 categories on the 6-point scale is shown in Figure 2. Compared with placebo, significantly greater proportions of patients in the golimumab groups showed an improvement of ≥2 categories at week 14 for the questions about trouble falling asleep, trouble staying asleep, and waking up feeling tired or worn out (Figure 2A). There was no difference between the groups in the proportion of patients with an improvement of ≥2 categories for the question about waking up several times during the night. The statistical significance of some of these comparisons was not maintained at week 24 (Figure 2B).
Changes from baseline to weeks 14 and 24 in JSEQ scores significantly correlated with changes from baseline in SF-36 summary scores, BASFI scores, night back pain, total back pain, BASDAI scores, and inflammation (morning stiffness) (Table 3). The strength of each correlation at week 24 was similar to the value at week 14. The correlations between changes in JSEQ scores and changes in PCS, BASFI, night back pain, total back pain, BASDAI, and inflammation (morning stiffness) were stronger than those between changes in JSEQ and changes in MCS. As expected, there were also strong correlations between changes in JSEQ scores, changes in the vitality component of the MCS score of the SF-36, and changes in the score from the fatigue question of the BASDAI.
In a multiple regression analysis that included changes in JSEQ score as the dependent variable and changes in PCS, MCS, BASFI, night back pain, total back pain, and inflammation (average of the last 2 questions of the BASDAI regarding morning stiffness) scores and treatment group as the independent variables, improvements from baseline to week 14 in PCS, MCS, and night back pain were significant predictors of improvements in JSEQ at week 14. At week 24, improvements from baseline in MCS, inflammation, and night back pain scores were significant predictors of improvements in JSEQ score (Table 4).
|Week 14 (r2 = 0.211, P < 0.001)|
|Change from baseline|
|Night back pain||0.274||0.003|
|Week 24 (r2 = 0.204, P < 0.001)|
|Change from baseline|
|Inflammation (morning stiffness)||0.253||0.013|
|Night back pain||0.358||0.0003|
The GO-RAISE study was, to our knowledge, the first randomized, placebo-controlled study that prospectively evaluated the effect of an anti–tumor necrosis factor α agent on sleep disturbance in patients with AS. Patients in the study had significantly disturbed sleep at baseline, with mean baseline JSEQ scores that represent ∼8 days of sleep disturbance in the previous 30 days (assuming each question contributes equally to the total score). The Cronbach's alpha coefficient for the JSEQ was 0.72, which was similar to the value reported in previous studies that used the JSEQ (13, 19). The JSEQ scores also correlated with other validated, patient-reported outcomes. These findings indicate that the JSEQ is a valid assessment of sleep disturbance in patients with AS.
The results of the current study showed that patients who received golimumab had significantly reduced sleep disturbance after 24 weeks in comparison with those who received placebo. The effect was similar between the golimumab dose groups and was only evident among patients with an ASAS20 response.
Golimumab-treated patients showed greater improvement in each individual question of the JSEQ, as shown by the proportions of patients with improvements from baseline of ≥2 categories on the 6-point scale. Each point on the scale is a categorical representation of the number of days with disturbance. Thus, a 2-category improvement does not represent a specific number of days. However, any patient who demonstrates this level of change would have experienced a significant reduction in sleep disturbance. At minimum, an improvement from a score of 2 to a score of 0 would represent a reduction from 4–7 days with disturbance to 0 days with disturbance. At maximum, an improvement from a score of 5 to a score of 3 would represent a reduction from 22–30 days with disturbance to 8–14 days with disturbance.
We observed statistically significant and clinically relevant correlations between improvements in JSEQ scores and improvements in SF-36, BASFI, night back pain, BASDAI, and total back pain scores, as well as in measures of fatigue from the BASDAI and vitality from the SF-36. In a multiple regression model, 14-week data showed that improvements in PCS, MCS, and night back pain scores were significant predictors of improvements in JSEQ score. Twenty-four–week data showed that improvements in MCS, inflammation, and night back pain scores were significant predictors of improvement in JSEQ score. Although MCS was a predictor of JSEQ improvement at both weeks 14 and 24, its contribution to each model was weak, supporting the notion that physical aspects of the disease are more important than mental or social aspects for sleep disturbance in patients with AS. Improvement in night back pain was the most significant predictor of improvement in sleep disturbance as measured by the JSEQ. Treatment group assignment was not a predictor of improvement in JSEQ score in the final model. However, each of these outcomes was significantly influenced by treatment group assignment. For example, median changes from baseline in night back pain were −0.3 in the placebo group, −3.0 in the 50-mg group, and −3.1 in the 100-mg group (P < 0.001 for all comparisons of golimumab versus placebo) (12).
In a previous study, Hultgren et al (6) evaluated sleep disturbance in outpatients with AS using the Uppsala Sleep Inventory questionnaire, which consists of 89 questions pertaining to sleep-related issues. The results showed that sleep disturbance was closely related to pain at bedtime and during the night. In our study, we did not evaluate pain at bedtime, but our results showing that night back pain was a significant predictor of sleep disturbance are otherwise consistent with those of Hultgren et al. In the recent Turkish study by Günaydin et al, disease activity was the most significant predictor of fatigue in a multiple regression analysis that included demographic, disease activity, and HRQOL variables, including sleep disturbance (10). However, this model differed from ours in that it evaluated fatigue as the dependent variable with sleep disturbance as an independent variable. In our model, sleep disturbance was the dependent variable.
In summary, patients with active AS showed significant sleep disturbance at baseline due to underlying pain associated with AS. Treatment with subcutaneous golimumab every 4 weeks significantly reduced sleep disturbance and improved HRQOL.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Deodhar 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 conception and design. Deodhar, Braun, Inman, Mack, Parasuraman, Buchanan, Hsu, van der Heijde.
Acquisition of data. Deodhar, Braun, Inman, van der Heijde.
Analysis and interpretation of data. Deodhar, Braun, Inman, Mack, Parasuraman, Buchanan, Hsu, Gathany, van der Heijde.
Employees of the study sponsors were involved in designing the study. Employees of the study sponsors, including Authors Mack, Parasuraman, Buchanan, Hsu, and Gathany, were involved in data analysis. Scott Newcomer, an employee of Centocor Ortho Biotech, assisted with the preparation of the manuscript. All authors reviewed and approved the manuscript for publication. Dr. Deodhar made the final decision to submit the manuscript.
We thank the patients, investigators, and study personnel who made the GO-RAISE study possible. We acknowledge Scott Newcomer, MS, of Centocor Ortho Biotech for his assistance in preparing the manuscript.