Patients with systemic lupus erythematosus (SLE) often have concomitant fibromyalgia (FM) or symptoms similar to FM, including chronic pain, fatigue, sleep disturbance, and/or depression, which are potentially distinct from immune-mediated disease activity and comprise Type 2 SLE symptoms,(1) “a prototypical nociplastic pain syndrome consisting of widespread pain in four of five specific body regions, with concurrent symptoms, including fatigue, sleep disturbance, memory problems, lack of concentration, and depression.”(2) Recent work showed that about one in six patients with SLE fulfill the 2016 classification criteria for FM(2, 3) (i.e., concurrent fibromyalgia), while nearly the same proportion reported chronic pain that does not qualify as a FM diagnosis such as osteoarthritis, and 65% had no pain.(4)

Changes in pain, co-occurring symptoms, and functional status are associated with rheumatologic conditions, making measurement of validated patient-reported outcome measures (PROs) critical for tailoring care and understanding trends in quality of life and associated clinical factors.(5) However, studies of SLE and FM vary by PRO measures, making comparability challenging.(6–10) Capturing and analyzing PROs can help identify potential interventions to improve quality of life, disease management, and survival in patients with rheumatologic conditions. In particular, manifestations of lupus affect physical function, fatigue, pain, and other PROs,(11, 12) potentially affecting survival.(13, 14)

This study explored patterns of symptoms using Patient-Reported Outcomes Measurement Information System (PROMIS)(15) measures to more deeply phenotype symptomatology, function, and quality of life in patients with SLE who may or may not have FM, particularly to expand on assessments with the 2016 American College of Rheumatology (ACR) FM criteria. This approach builds on previously published studies that sought to discern typologies of SLE that extend beyond classical clinical findings.(1, 16, 17) The goal was to understand how PROMIS measures might vary by pain group and whether there might be additional value of wider score ranges and more granular scores for disease identification, characterization, and management, including whether PROMIS measures might serve as potential alternatives to classify symptom clusters in a population with SLE who may or may not have laboratory results or physician assessment indicative of SLE activity. This was approached by comparing patient-reported responses to the FM criteria survey around the same time as the NIH PROMIS PROs in a cohort of SLE patients seen in an outpatient setting.

The final sample of 181 patients came from the 316 patients who completed the FM questionnaire who also completed at least one PROMIS measure (57%). The sample was 89% female, 34% Black/African American, 19% Asian, 29% white, and 23% Hispanic, and with a median age of 39 years. Of the total 181 patients, 30% (n=54) had chronic pain “more days than not during the last 3 months” and continued to complete the FM questionnaire (Supplementary figure 1). The cohort included 31 (17%) with FM, 23 (13%) with non-FM chronic pain, and 127 (70%) with no chronic pain.

There were no significant differences by pain group in demographic characteristics, days between PROMIS and FM survey administration, SLEDAI score, or SLICC damage index score. However, of those in DORIS remission, there were proportionately more persons in the no pain group (p=0.017); DORIS remission was achieved in 32% (n=40/127) of those without chronic pain versus 13% (n=4/31) and 17% (n=4/23) in those with FM and non-FM chronic pain, respectively.(Table 2) The absolute number of days between FM survey and PROMIS responses showed no significant differences in PROMIS scores by days between surveys (mean 1.4 days, median 0 days; p>0.13, data not shown).

Overall, FM Criteria Survey questions on fatigue and memory issues were significantly different between the two chronic pain groups (i.e., categorized by the FM Criteria Survey as those with chronic pain, with or without fibromyalgia, n=54)(p<0.02). PROMIS scores for cognitive function were similar for the two chronic pain groups (41.2 for FM pain group and 41.4 for non-FM pain), while there appears to be a trend toward more severe “memory issues” in the FM pain group (74% of FM group ranking a 2 or 3 on memory issues and 39% of non-FM chronic pain group).(Table 2)

When comparing all three groups (n=181; FM, non-FM chronic pain, and no chronic pain) by PROMIS scores for fatigue, sleep disturbance, cognitive function, and depression, these measures were significantly different across the three groups (p<0.001), with the FM group consistently having higher symptom and lower function scores, the non-FM chronic pain group in the middle, and the no chronic pain group having the lowest symptoms and highest function scores. (See Table 2; Figure 1; Supplementary figure 2.) The pattern also held for PROMIS pain intensity and pain interference scores (p<0.001). Exploratory pairwise tests suggested that both FM and non-FM pain groups were significantly different from the no-pain group on all PROMIS measures. The FM and non-FM pain groups were not significantly different by PROMIS measures, except for PROMIS fatigue and pain interference scores.

The FM survey and PROMIS score response trends were concordant. Exploratory analyses suggested that among participants completing the FM survey, the PROMIS depression score median was higher at 60.5 for those reporting depression (i.e., moderate depression for PROMIS) versus 53 for those not reporting depression (p<0.001). Unlike the dichotomized FM response categories for depression, fatigue, sleep, and memory, measures on the FM survey ranged 0–3. As FM survey responses for fatigue worsened from 0–3, the PROMIS fatigue score also worsened (median=59 to median=68, p<0.001); those rating a 2 or 3 on the FM survey reported moderate fatigue according to PROMIS scores. As FM survey responses for “waking unrefreshed” worsened from 0–3, the PROMIS sleep disturbance score also worsened (median=54 to median=64, p<0.001); PROMIS scores indicated moderate sleep disturbance for FM survey responses of 2–3. Finally, as the FM survey response for “memory issues/lack of concentration” worsened from 0–3, PROMIS cognitive function scores worsened (median=48 to median=36, p<0.001) from normal to moderate symptoms according to PROMIS.

As demonstrated here, the overlapping PROMIS patient-reported measures for pain, depression, fatigue, sleep disturbance, and cognitive function are strongly associated with the 2016 FM criteria survey questions and show statistically significant differences among patients by pain group and directional trends as might be expected. Using a disease-independent PRO measurement tool, such as PROMIS, would allow for exploration and definition of patient subtypes across a wider array of measures than the FM Criteria Survey, potential development of a screening tool, advancement of FM research, and further study of the impact of medications and other interventions for patients with rheumatic conditions and secondary FM, a highly prevalent condition in patients with rheumatic diseases.(2, 3)

In SLE, there has been a recognition of two classes of symptoms: Type 1 or typical inflammatory symptoms, and Type 2 or symptoms more closely related to FM.(1, 16, 17) In most studies of Type 2 SLE, there is differentiation between a patient who meets concurrent FM criteria or has isolated and infrequent Type 2 symptoms. As there may be two or more separate, underlying processes occurring, it would be helpful to define these patient groups as also meeting FM criteria or having Type 2 symptoms without FM. This cross-sectional study revealed Type 2 lupus symptoms exist in patients with SLE who report chronic pain, regardless of whether they meet criteria for concurrent FM. These symptoms are not associated with SLEDAI activity or SLICC damage score and may represent a distinct entity, as others have suggested.(5) While SLE patients with FM have more prominent Type 2 symptoms, the presence of similar symptoms in the non-FM chronic pain group raises important questions regarding pathophysiology. It may be beneficial to study these groups separately as it is not clear if the Type 2 symptoms in lupus patients with non-FM chronic pain and lupus patients with FM represent two or more separate underlying processes. Other work has shown PROMIS scores can be used to identify distinct patterns of Type 2 symptoms in patients with SLE (5); our study expands on that to explore trends in scores for other PROMIS measures in a SLE cohort that includes diagnosed FM patients; adding measures may help identify other classes of symptoms/function typologies in future work.

Our study shows that SLE patients with non-FM chronic pain also have symptoms common to FM, such as fatigue, sleep disturbance, and cognitive issues, more often than SLE patients without chronic pain, but they do not have these symptoms as severely as those meeting FM criteria. Studies have been addressing the optimal tool to measure Type 2 symptoms for research purposes.(5, 16) We propose PROMIS measures as an appropriate instrument to gauge and address Type 2 SLE, because PROMIS measures show more variation and nuance between SLE patients in pain with and without FM; further, computer-adaptive testing tools make survey completion quicker and tailored to each patient.(24) It has been postulated that FM can develop in patients with inflammatory disease that is not well controlled.(4) That our patients in DORIS remission were more likely to be in the no-pain group supports this hypothesis. However, the most interesting group to be explored longitudinally are those in DORIS remission who have FM. Patients may have different underlying pathophysiology, and it will be important to distinguish between these when studying treatment options and symptom changes.

This study has limitations, including recall period differences, definitional differences, and small cell sizes for some responses. Both the FM survey and PROMIS measures are point-in-time surveys that have different recall periods (i.e., in the last week vs. the last 3 months). The FM survey asks questions with yes/no or 0–3 response ranges, while PROMIS yields scores ranging from 20–80. In particular, a yes/no response on the FM survey gives no information about severity (e.g., mild, moderate, severe); in some cases, a “no” or “0” response may be due to the absence of a formal diagnosis (e.g., depression). The PROMIS pain measures are conceptually different from those in FM, with the latter primarily focused on the number of body parts and locations that hurt persistently over 3 months while the former asks about pain interference and intensity. Of note, small numbers may contribute in part to observed differences identified between groups and make robust conclusions regarding associations between FM survey response and PROMIS scores challenging. The study environment yielded a broadly representative cohort of patients who have SLE in an urban setting.

To further understand differences in disease manifestation and patient experiences of their health condition that may be reflective of differential pain and disease processes and treatment effects, study results suggest integrating externally-validated PROs that are sensitive to changes in health status, such as PROMIS measures. Future work should incorporate PROMIS measures into studies to understand nuanced trends and symptom severity over time, because such measures can highlight meaningful and quantifiable changes in outcomes(21) and potentially identify new, more complete symptom classification that goes beyond laboratory results or physician classification.