Introduction

Systemic lupus erythematosus (SLE) is a multisystemic chronic autoimmune disease with an unpredictable waxing and waning course. Although survival has improved dramatically during the past few decades, life expectancy remains shorter and quality of life poorer than in the general population.

SLE may profoundly affect not only the physical aspects of a person’s life but also his or her mental, social, psychological, and sexual well-being, none of which are captured by current measures of disease activity or damage. Disease activity is assessed by different validated instruments: the Systemic Lupus Activity Measure-Revised (SLAM-R); the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) and its revised version, the SLEDAI-2000; and the British Isles Lupus Assessment Group Index. Accrued damage, on the other hand, is ascertained by the Systemic Lupus International Collaborating Clinics Damage Index (SDI) [1–4]. The application of these instruments requires skillful and trained physicians; only the SLAM-R incorporates the patient’s perspective on disease activity.

In the recent past, the Outcome Measures in Rheumatology Clinical Trials (OMERACT) group and the US Food and Drug Administration have recommended also assessing the patient’s experience with the disease through health-related quality of life (HRQoL) or patient-reported outcome measures; this applies to both clinical trials and longitudinal observational studies [5, 6]. These measures reflect the individual’s perspective on the impact of a disease and its treatments on his or her function and perceived well-being, including the physical, mental, and social domains of life at a single time point [7]. These measures also provide an important facet of the treatment response in clinical trials and daily clinical practice and may aid in the formulation of specific treatment recommendations.

Health-Related Quality of Life

Generic Measures

Generic measures are commonly used among patients with different diseases or conditions and in various populations and clinical settings. They are useful to compare populations with different disorders of varied clinical severity as well as to compare patients with a given condition with healthy individuals. They lack specificity and are not useful to identify certain domains that may be especially important to SLE patients, such as sexual functioning and body image. The Medical Outcomes Survey Short Form 36 and 20 (SF-36 and SF-20) have been the most commonly used in multiethnic, multicenter, and single-center studies [8••, 9–14••]. Other generic measures, such as the European-QoL scale (EQ-5D) and the World Health Organization-QoL Scale (WHOQoL-Bref), the Health Assessment Questionnaire (HAQ), the Arthritis Impact Measurement Scale (AIMS), the Sickness Impact Profile, and the quality-of-life scale, have been used in SLE, but not in multiethnic SLE cohorts [15–19]. The HAQ and AIMS, as measures of functional status, have not been universally accepted in SLE research despite the frequency of neurological and musculoskeletal manifestations in SLE patients.

SLE-Specific Measures

These include domains not addressed by generic measures and identify some patient characteristics that are not usually captured by the current biomedical model, permitting the distinction between useful and useless therapeutic interventions in SLE clinical trials and in clinical care settings. Disease-specific measures were developed to capture the patient’s perspective on the impact of SLE and its treatments over time in a sensitive and reliable manner. These measures are the following: the Systemic Lupus Erythematosus–Specific QoL (SLE-QoL), the SLE-QoL questionnaire (L-QoL), the Lupus-QoL Scale (LupusQoL), the LupusPRO, the Lupus-QoL (LUP-QoL), and the Simple Measure of Impact of Lupus Erythematosus in Youngsters (SMILEY) [20, 21•, 22•, 23••, 24•, 25, 26••, 27].

To adapt a new quality-of-life (QoL) instrument across cultures and preserve the equivalence of what it measures, these steps should be followed: preparation, forward translation, reconciliation, back translation and its review, harmonization, cognitive debriefing and its review, results, finalization (expert judgment), proofreading, pretesting (probe technique or testing the instrument in bilingual laypeople), weighting of scores (ie, adapting the weights of scores to the cultural setting), and final report of the QoL instrument [28]. Given that most clinical studies are conducted in Western European or North American cultural environments (at least until recently), the cultural adaptation of HRQoL measures is very important to avoid selection bias associated with studies that might selectively exclude non-English-speaking patients [28]. In addition, the measurement properties of the adapted HRQoL instrument should be known and demonstrated by its use in the target population. The preservation of the metric properties of the HRQoL instrument must take into account internal consistency (measures the correlation between answers to different questions about the same concept) and reliability or repeatability. A measure is reliable at the group level in terms of internal consistency if Cronbach’s α is greater than 0.7. The test–retest reliability of the measure assesses its stability over time and should be satisfactory.

Validation studies are required when implementing a new HRQoL measure. In addition to feasibility, all HRQoL instruments should demonstrate content validity (measures what is intended to measure), discriminant validity (distinguishes disease severity, captured by disease activity or damage), criterion validity (measures accurately the same phenomenon), construct validity (measures accurately the underlying construct), convergent validity (measures the extent of correlation between observed relationships of the concepts and hypothesized concepts), sensitivity to change or responsiveness (detects changes individually or in groups of people), and finally the minimal clinically important difference (clinically important/relevant change in disease status) [28, 29••]. Compliance with these steps allows the comparison of dissimilar health care systems across and within countries and maintains internal and external validity of the HRQoL instrument.

Validated SLE-Specific Measures

The salient characteristics of SLE-specific QoL measures are depicted in Table 1. The SLE-QoL has been developed and validated by Leong et al. [20] in SLE patients from Singapore; it was found to be more sensitive, but not specific, to change than the SF-36, although its concurrent validity remains an issue and the sensitivity to change in other populations remains to be determined [20]. The instrument was recently cross-culturally adapted for Chinese-speaking SLE patients from Singapore [30].

Table 1 Lupus-specific health-related quality-of-life instruments
Full size table

The L-QoL instrument, developed by Doward et al. [21•], was derived from in-depth interviews with SLE patients and measures the impact of the disease and its treatments; however, most patients were in fair to excellent health. It has good face and content validities as per cognitive debriefing conducted with United Kingdom SLE patients, but it requires larger validation studies. Hungarian and Turkish versions of the L-QoL are currently available. Sensitivity to change remains to be proven [21•].

The LupusQoL, developed by McElhone et al. [22•], was derived from semistructured interviews with predominantly Caucasian SLE women from the United Kingdom. The relevance of the items included in the instrument was confirmed by cognitive debriefing [22•]. The LupusQoL has been assessed in United Kingdom SLE patients and demonstrated good internal consistency, test–retest reliability, and concurrent validity with the SF-36. It has also shown good discriminant validity for different levels of disease activity and damage. The LupusQoL was recently adapted for cross-cultural use and adapted, validated, and evaluated for use in the United States (LupusQoL-US), but its sensitivity to change has yet to be determined [23••]. It has been adapted to Canadian-English and translated into other languages, including Spanish, Dutch, French, Greek, Hungarian, Italian, Portuguese, Swedish, and Chinese. Cross-cultural validation of this tool in these other languages is under way.

The LupusPRO is a disease-targeted, patient-reported outcome measure that has been developed in SLE patients (female and male) of heterogeneous ethnicity within the United States. It has HRQoL and non-HRQoL items and may be used with a generic HRQoL tool in clinical research. It has good content validity, internal consistency reliability, test–retest reliability, and construct and convergent validity against the SF-36, EQ-5D, and LupusQoL-US [31]. It is responsive to changes over time. A Spanish version is being validated in Hispanic SLE patients from the continental United States, Puerto Rico, Mexico, and South America. Adaptation/translation and cross-cultural validation for Canadian-English, Canadian-French, Turkish, and other languages as well as responsiveness are currently being evaluated [24•].

The LUP-QOL measure developed by Yu et al. [25] was found to be a valid and reliable instrument in 121 SLE patients with mild to moderate disease, in whom it showed high test–retest reliability and validity.

Moorthy et al. [26••, 27] developed the SMILEY, a brief and valid English-US HRQoL instrument for children and adolescents with SLE (≤19 years of age) that is currently being cross-culturally adapted, validated, and having its responsiveness determined in many other languages and countries, including Danish, Dutch, French (France), German (Germany), Hebrew, Italian, Portuguese (Brazil), Slovene, Spanish (Spain, Argentina, Mexico, United States, and Puerto Rico), and Turkish.

Cross-Cultural Adaptation of QoL Instruments

Patients from some ethnic and racial groups tend to be younger at disease onset, to have a more severe disease course, and to have worse outcomes (eg, more acute onset or more severe lupus, including lupus nephritis [LN], in African Americans and Hispanics) [32]. The cultural features of SLE patients from different ethnic backgrounds affect the course of the disease; health behaviors, beliefs, and medical care systems vary by ethnicity, cultures, and nations. For example, in developing nations, SLE patients and those with other chronic diseases tend to seek medical care only when life-threatening symptoms occur or when disability is present. This phenomenon also has been reported among minority groups in the United States. In addition, it is well-known that socioeconomic status (SES; poverty is an SES marker) influences profoundly the course of the disease and its outcomes [33•]. These issues may contribute to unique QoL concerns and health outcomes among SLE patients based on ethnic/cultural differences. Thus, SLE is a disease with protean manifestations that leads to different responses in patients (biologically, psychosocially, and culturally), health professionals, and society itself, giving rise to different health policies and health care provisions. As SLE predominantly affects younger patients, they inherently have a greater number of years of productive life at risk as compared with patients with other chronic diseases. Also, they may not have acquired significant social or economic resources due to their relative young age. The social, emotional, and economic burdens imposed by SLE may in addition become significant QoL concerns when the cost of long-term medical care is primarily a patient’s own responsibility. These issues affect the generalizability of various SLE-specific QoL tools in various ethnic groups or medical care settings.

The emergence of multinational, multicenter studies in SLE underscores the need for validated HRQoL measures to compare patients in an equivalent manner across and within nations, taking into account not only different languages and dialects but also cultural differences among these patients. Obviously, spoken languages and dialects and how they are used vary within and among countries, affecting the adaptation/validation and the subsequent understanding of HRQoL instruments. To make studies comparable, equivalent use of HRQoL instruments is essential. To achieve this goal, the cross-cultural use of HRQoL outcome measures requires careful cross-cultural adaptation (to capture the same concept in another cultural setting equivalently), as well as validation of the transformed instrument [28]. Despite the cross-cultural validation of an existent tool, whether a tool developed in a particular ethnic or cultural background will be contextually comprehensive in content validity for another ethnicity/culture/medical care system remains to be determined.

Observational Studies Assessing the Relationship Between Clinical and Nonclinical Variables with HRQoL in SLE Populations with Different Sociodemographic Features, Ethnic Background, and Disease Status

The most relevant characteristics of selected observational studies using HRQoL measures in SLE are shown in Table 2.

Table 2 Selected observational studies using HRQoL measures in SLE
Full size table

The SF-36 has been used successfully among SLE patients from diverse Western societies (United States, Canada, Norway, and Spain) and in other settings, such as Singapore, where it has been shown to have satisfactory psychometric properties and construct validity. The instrument has been found to be suitable for different sociocultural environments and has been used in cross-sectional and longitudinal studies. Furthermore, it has been shown that reductions in disease activity are associated with better HRQoL [8••]. However, in a longitudinal study of 146 University of Toronto Lupus Clinic patients (predominantly Caucasian [75%] and female [90%]), disease features (disease activity, steroids, or damage accrual) were not associated with changes in physical functioning, except for the presence of fibromyalgia [34••]. It is likely that the ethnic composition (Caucasian predominance), longstanding disease, the presence of mild to moderate disease activity, and these patients’ low SDI explain the lack of sensitivity to change of the SF-36 over an 8-year period. This study also highlights the potential usefulness of documenting fibromyalgia and/or pain/tenderness in SLE patients.

In another predominantly Caucasian SLE cohort (n = 386) from Montreal, Canada, the relationship between renal activity and HRQoL (assessed by the SF-36) was studied because it has been reported that irreversible renal damage does not influence HRQoL [35]. SLE patients with concurrent active renal disease, as defined by the SLEDAI renal items, experienced a slightly poorer HRQoL—especially in the physical domains—than those without renal disease; however, in this study, changes in renal activity associated with changes in QoL could not be estimated [36••].

In contrast to the above studies, those that have included SLE patients with an ethnic composition associated with a potentially more serious disease phenotype have shown discrepant results. For example, the effects of disease activity and damage on the SF-36 were assessed in age- and sex-matched Chinese SLE patients (n = 155) observed longitudinally for 2 years with repeated HRQoL measurements; previous organ damage was associated with poorer HRQoL, and new damage predicted a further decline in HRQoL, whereas persistent disease activity was associated with deterioration in some subscales of the SF-36 [37•].

Demographic and HRQoL variables in two different African American patient groups, one urban from Chicago (n = 137 with high genetic admixture [20–30%]) and the other rural from South Carolina (n = 35 with low genetic admixture [9.8–12%]) were compared using the SF-36. In this predominantly female study, HRQoL was significantly worse in urban African Americans than in the age- and gender-matched rural patients [38•].

In the LUMINA (Lupus in Minorities, Nature Versus Nurture Multiethnic Cohort; 19% Hispanic-Texans, 17% Hispanic-Puerto Ricans, 35% African Americans, 29% Caucasians), the SF-36–derived utility measure or SF-6D was assessed in 2662 visits from 588 SLE patients, and all variables from the enrollment or preceding visits were examined to predict HRQoL. A single number represents health status in the SF-6D; the lower the value, the poorer the QoL is. Within this multiethnic cohort, Hispanic-Texan ethnicity and higher levels of social support predicted HRQoL, whereas other variables known to affect disease outcomes (older age, poverty, greater disease activity and damage), as well as fatigue, helplessness, and abnormal illness-related behaviors were negative predictors, with prior SF-6D being the strongest predictor of subsequent HRQoL [39••].

Whether SF-6D measured early in the disease course is associated with damage accrual and mortality was also examined in 552 patients from the same cohort. SF-6D was associated with damage accrual, but not with mortality. The physical component summary scale (PCS) and mental component summary scale (MCS) of the SF-36 were also examined; the PCS, but not the MCS, was found to be associated with damage, but not with mortality [40••]. The SF-6D originally had been derived for economic evaluations as a utility measure and has been shown to be a reliable instrument to assess HRQoL. Especially during early disease, the SF-6D seems more sensitive than the SF-36 to detect subtle changes in HRQoL and more likely to predict damage accrual. In SLE, mortality is related to SES factors such as age, ethnicity, and poverty, as well as to disease activity and damage; in the LUMINA study, neither the SF-6D nor the PCS and MCS scores of the SF-36 predicted mortality. However, it is possible that the influence of SF-6D in mortality is mediated by damage accrual. Until now, the role of HRQoL measures as a predictor of mortality in SLE has not been demonstrated in large cohort studies. Nevertheless, a small study conducted in 63 Brazilian SLE patients suggested that the most important predictor of mortality at 5 years is a mean low score in the role emotional subscale of the SF-36. As expected, SDI scores were higher in those who died [41].

Wolfe et al. [42••] have shown that HRQoL was predicted by disease damage, comorbidity, and SES features (income, education, and age), and that overall EQ-5D and PCS scores were at the same levels in SLE as in RA and noninflammatory rheumatic diseases, but the MCS and mental health subscale were more abnormal in SLE.

Satisfactory psychometric properties of the EQ-5D and SF-6D were recently reported by Aggarwal et al. [14••] in a multiethnic US SLE cohort from Chicago; however, these measures were weakly correlated with disease activity and damage [14••].

Khanna et al. [43] assessed QoL using the WHOQoL-Bref instrument in 73 mostly female Indian (n = 70) patients with short disease duration and showed that higher disease activity scores were also associated with lower QoL scores in the physical and psychological domains.

It is well-established that cutaneous lupus erythematosus (CLE) may have a negative impact on SLE patients’ QoL, particularly when it affects body image. Recently, the relationship between CLE and QoL was studied in 157 SLE patients in whom the QoL of patients with CLE was compared with the QoL of patients with other skin diseases. QoL was assessed using the Skindex-29 (symptoms, emotions, and functioning, with higher scores indicating worse QoL) and disease severity with the Cutaneous Lupus Erythematosus Disease Area and Severity Index. In this study, CLE had a negative impact on QoL, especially in women with more severe CLE [44]. Similar findings have been reported for body image and HRQoL in SLE by Jolly et al. [45].

The findings from the above-mentioned studies stress the role and importance of ethnic composition (including genetic admixture and gene expression), disease duration (early vs late disease), the underlying biology of SLE patients (disease phenotype), the setting in which the HRQoL instruments are administered, as well as the importance of SES, including behavioral factors (eg, smoking, compliance), and emotional aspects, including fibromyalgia, in patients’ self-perception of QoL. Several studies have also shown a lack of association between disease activity and generic HRQoL measures, especially with SF-36, which suggests that generic HRQoL assessment tools may not be the most sensitive to assess HRQoL over time [13]. Lack of correlation between self-perceived patient health status and clinical symptoms is well-known (eg, patients with active LN tend to underestimate their overall health in visual analogue scales [VAS], and fibromyalgia patients tend to overestimate it). Thus, it is conceivable that HRQoL measures are more useful and sensitive to signals of clinical improvement rather than to those pointing out clinical deterioration. Even though in cross-sectional studies, poor correlation may be noted between HRQoL and disease activity, it is plausible that in longitudinal studies (eg, clinical intervention trials or larger observational studies), evaluation of within- and between-patient variations in health status and disease activity over time in response to the intervention or natural waxing/waning disease course may be better appreciated

Therapeutic and Nontherapeutic Interventions Influencing HRQoL in SLE

In the lupus community, the ability to assess interventions aimed at controlling not only potentially life-threatening complications (eg, severe LN, neuropsychiatric involvement) but also mild to moderate SLE manifestations should not be underestimated. The role and impact of several therapeutic interventions on self-perceived HRQoL have been examined in open-label studies and in randomized controlled clinical trials of pharmacologic and nonpharmacologic interventions.

The Dutch LN Study Group compared the treatment effect of cyclophosphamide pulses plus oral prednisone versus azathioprine combined with intravenous methylprednisolone during a 24-month period on HRQoL in 87 (predominantly Caucasian) patients with proliferative LN. HRQoL and disease activity were measured at baseline and at 12 and 24 months using patient VAS, SF-36, profile of mood states, and the SLE Symptom Check List [46]. The treatment burden was assessed at 24 months, and disease activity was measured with the SLEDAI and the physician’s VAS. Complete questionnaires were available in 47 patients. The study showed that aggressive treatment of LN with immunosuppressive drugs and concomitant glucocorticoids improved HRQoL, especially in the first year [47••]. In keeping with these results, the randomized controlled trial conducted by Dussán et al. [48••] compared high-dose cyclophosphamide (50 mg/kg for 4 days) with monthly intravenous cyclophosphamide in first-time cyclophosphamide users, demonstrating improvements on several components of the SF-36 measures at 6 months in patients receiving high-dose cyclophosphamide, but equal improvement in the two arms at 36 months.

Similarly, the effect of cyclophosphamide pulse therapy on HRQoL was examined in a Brazilian study that compared cyclophosphamide users with nonusers to ascertain treatment burden. In this study, cyclophosphamide pulse therapy did not worsen HRQoL; rather, psychological distress was the main explanatory variable of poor HRQoL in these patients [49].

A small open-label study of the biological agent infliximab at conventional doses and schedule showed that infliximab improves SLEDAI without altering any of the SF-36 subscales [50].

Hydroxychloroquine prevents lupus flares and accrual of disease damage while improving patient survival. Thus, its effects on HRQoL were studied cross-sectionally in 230 SLE patients (60% African American, 20% Caucasian, 14% Hispanic, 6% Asian) from one Chicago area center using the LupusQoL-US. As expected, hydroxychloroquine use was associated with decreased damage accrual; however, HRQoL status was apparently not affected when using the LupusQoL-US or the LupusPRO measure to ascertain it [51••].

These data suggest that aggressive interventions aimed at obtaining rapid disease control during active disease are not perceived by patients as a treatment burden. However, patient perception may change over time with subsequent treatment exposure when the long-term side effects become apparent and disease damage ensues. Given that hydroxychloroquine is not indicated for rapid, active disease control, it is therefore reasonable to expect its long-term benefits in retarding the occurrence of damage and on survival are not perceived by patients, as this was already corroborated in a well-conducted observational study [52].

Other therapeutic interventions also may improve the QoL of SLE patients. For example, hip arthroplasties contributed to improved HRQoL as measured by the SF-36 in a group of Japanese patients [53].

Not every therapeutic intervention may prove useful and safe; in fact, the use of complementary alternative medicine in SLE was associated with greater cumulative disease damage and worse QoL; however, some other nontherapeutic interventions, such as patient education, may be beneficial to patients’ QoL [54, 55].

Conclusions and Research Agenda

HRQoL has been assessed in SLE populations with different degrees of severity. Studies have included mostly patients from Western societies and very few from developing societies, with small sample sizes. It is still not known how different self-perceived HRQoL may be among men with SLE, as most studies have been conducted in women, who make up the majority of SLE patients. Overall, HRQoL is poor in SLE patients regardless of ethnicity, culture, geographic location, or SES; the factors that have emerged as being implicated in yielding a poor HRQoL experience in SLE are sociodemographic (age, poverty, education) and disease related (higher disease activity and damage, fibromyalgia, fatigue, helplessness, abnormal illness-related behaviors, and mental health domains).

SF-36 has been the most commonly used HRQoL assessment tool; however, it may not be the optimal instrument to ascertain HRQoL over time given its poor sensitivity to change. However lupus-specific HRQoL measures have not yet been used in any SLE longitudinal study.

The SF-6D independently predicts disease damage, but not mortality, especially in early disease.

A low SES may influence HRQoL perception and its ascertainment, and it may alter biological and nonbiological manifestations, such as pain perception. Thus, SES should be included in any new study of HRQoL and carefully ascertained using appropriate tools and statistical models.

Fibromyalgia, a common chronic comorbid condition that is present worldwide, may distort health perception in any sociocultural and geographic setting and contribute to poor HRQoL; it should be carefully documented, included in HRQoL studies, and addressed in the clinical setting.

Aggressive therapeutic interventions, either educational or pharmacologic, especially with immunosuppressive agents aimed to control active, severe lupus manifestations are well-perceived by patients and physicians.

Medical interventions directed at controlling grumbling disease and preventing known long-term deadly complications, such as atherosclerotic cardiovascular disease, may not be fully appreciated by SLE patients or even by some physicians.

The use of HRQoL research tools such as the LupusPRO and the SMILEY, which are currently undergoing validation studies in different countries, languages, and cultural settings, may help dissect the underlying role of SES and disease-related features not only in QoL but also in other intermediate and long-term outcomes (disease-related damage and mortality) and will help to overcome the limitations posed by ethnic heterogeneity within a given lupus cohort or between lupus cohorts with dissimilar ethnicities.