Differences in patient-reported outcomes (PROs) by disease severity in light chain (AL) amyloidosis

Abstract

Objective:

To assess the impact of organ involvement on patient-reported outcomes (PROs) in light chain (AL) amyloidosis.

Methods:

PROs were evaluated using the KCCQ-12, PROMIS-29+2, and SF-36 in individuals with AL amyloidosis. The 2004 Mayo system was used to stage disease, and cardiac, neurologic, and renal involvement was considered. Global physical and mental health (MH) scores, physical function (PF), fatigue, social function (SF), pain, sleep, and MH domains were evaluated. Effect sizes between scores were measured using Cohen’s d.

Results:

Of 297 respondents, the median age at diagnosis was 60 years with 58% cardiac, 58% renal, and 30% neurologic involvement. Fatigue, PF, SF, and global physical health with PROMIS and SF-36 discriminated the most by stage. Significant discrimination in PROMIS and/or SF-36 was seen in PF, fatigue, and global physical health with cardiac involvement. For neurologic involvement, PF, fatigue, SF, pain, sleep, global physical and MH with PROMIS and role physical, vitality, pain, general health, and physical component summary with SF-36 were discriminatory. For renal amyloid, pain by SF-36 and PROMIS, and SF-36 MH and role emotional subscales were significant.

Conclusions:

Fatigue, PF, SF, and global physical health can discriminate stage, cardiac and neurologic, but not renal, AL amyloidosis involvement.

INTRODUCTION

Light chain (AL) amyloidosis and its treatment are associated with impaired health-related quality of life.¹ This rare hematologic disease arises from a clonal plasma cell disorder, such as a monoclonal gammopathy of undetermined significance and multiple myeloma. The disease presents as a multisystemic disorder characterized by organ dysfunction due to deposition of insoluble fibril deposits.^{2, 3} The heart is the most commonly involved organ, seen in nearly 80% of patients, with development of a restrictive cardiomyopathy with diastolic dysfunction and eventually, heart failure.² The severity of cardiac involvement determines stage and dictates the prognosis of this disease.⁴ Renal, neurological, gastrointestinal, and soft tissue organ involvement can also occur, though the number and pattern of organs involved by amyloidosis is often heterogeneous. The treatment of AL amyloidosis primarily involves chemotherapy to control the underlying plasma cell clone, including autologous stem cell transplantation in eligible patients along with aggressive supportive care.² Currently, there are no approved fibril-directed treatments to remove pre-formed amyloid fibrils from organs. Thus, organ amyloidosis may never reverse fully. Patients with AL amyloidosis often experience high symptom burden associated with the type of organs affected by amyloid deposits, severity of organ dysfunction, as well as the treatment of the disease itself.^{5, 6}

There are multiple options available to measure patient-reported outcomes (PRO) in AL amyloidosis.^{7, 8} The SF-36 has been the most widely studied, with documented content validity ⁹ and acceptable psychometric properties in AL amyloidosis patients treated at an academic center and from a community-based sample.¹⁰ Some PROMIS scales have also shown evidence of internal consistency, reliability, and construct validity in small cohorts of newly diagnosed and established AL amyloidosis patients from academic centers.^11-13 The KCCQ-12 is a 12-item short form for use in patients with heart failure.¹⁴ The KCCQ-12 has been used in amyloidosis¹⁵ because cardiac involvement is common and determines prognosis. There remains a need for additional information on the performance of these measures in wider samples of AL amyloidosis patients, and no study has used all three measures in the same sample. We conducted the current study to compare the KCCQ-12, PROMIS, and SF-36 instruments in a community-based sample of AL amyloidosis patients. We evaluated known groups validity by comparing the 3 measures on how well they discriminate between patients by disease severity and pattern of organ involvement of AL amyloidosis.

METHODS

AL amyloidosis sample:

People living with AL amyloidosis who were members of the Amyloidosis Support Groups, Inc. (ASG) were invited to participate in an IRB-approved survey. The study invitation was disseminated by the President of ASG (MF) via email and shared on a closed Facebook group (AL amyloidosis ASG). Participants were provided an informational letter, and those who completed the online survey were compensated with a gift card of $50. The survey was open between 7/2/2021 to 7/28/2021.

In addition to baseline sociodemograhic information, individuals provided information on their amyloidosis, including time from diagnosis, type and number of organs involved, cardiac biomarkers at diagnosis and closest to survey completion, and treatment of their disease. The cardiac biomarkers NT-proBNP, BNP, troponin T (4^th generation and 5^th generation), and troponin I were considered.¹⁶ These data were reviewed (AD and AS) to assess plausibility and values that appeared erroneous were ignored. For example, a troponin T or NT proBNP value input as n/a, not measured, or inclusion of a date instead of a value were ignored. However, all other reported data from that patient was still used in the analysis. The 2004 Mayo Clinic stage⁴ was thus calculated in a subset of participants at diagnosis (N=106) and at time of survey (N=101). Stage at time of survey completion was analyzed.

Measure Description:

The KCCQ-12 is a 12-item questionnaire designed to measure several important aspects of heart failure.¹⁷ It includes a total index score and four subscale (physical limitation, symptoms, quality of life, and social limitation) scores. The scale is scored from 0 to 100 (higher scores = better health status). PROMIS items included Global health v1.2,¹⁸ PROMIS-29+2 v2.1 profile,¹⁹ and PROMIS Short Form v1.0 Fatigue-8a. The HealthMeasures Scoring Service was used to score each domain.²⁰ PROMIS scores are expressed as T-scores, for which a score of 50 corresponds to the US general population with a standard deviation (SD) of 10. Higher scores correspond to more of the domain (e.g., Physical function and social roles- score >50 denotes better physical function than general population average, Fatigue- score >50 denotes greater fatigue than general population average). Because of the prevalence and impact of fatigue in AL amyloidosis patients, we administered the Fatigue-8a short form which includes the 4 items on fatigue severity that are also part of the PROMIS-29 with the addition of 4 items asking about fatigue impact. Two summary scores, global physical health summary (GPHS) and global mental health summary (GMHS), and eight domains, physical function, fatigue, ability to perform social roles and activities, anxiety, depression, pain interference, sleep disturbance, and cognitive function, were scored. The Medical Outcomes Study Short Form (SF-36 v.1) is a licensed 36-item, generic measure of health-related quality of life.²¹ The scores are derived using a norm-based scoring strategy that yields standardized distributions with a mean of 50 and a SD of 10 in the US general population. A higher score implies better health status. Two summary scores, physical component scale (PCS) and mental component scale (MCS), and eight subscale scores including physical function (PF), bodily pain (BP), role limitations due to personal or emotional problems (RE), general mental health (MH), social functioning (SF), energy/fatigue or vitality (VT), and general health perceptions (GH) can be calculated. Scoring was done using instructions provided by QualityMetric.²² The Physical and Mental Component Scores were computed using the oblique rotation method of Farivar et al.²³

Clinical groups:

The following disease groupings were considered: 2004 AL stage I vs II vs III, cardiac organ involvement yes compared to no, neuropathic involvement yes compared to no, and renal involvement yes compared to no. We hypothesized that patients with stage I AL amyloidosis would have at least a 3-point difference in mean score with higher physical function and lower fatigue than stage III AL and that individuals with cardiac AL will have at least a 3-point difference in mean score for lower physical function and social roles, and higher fatigue compared to those with no cardiac AL. We did not have an a priori hypotheses for differences in PRO scores for renal and neuropathy known groups; these differences were considered exploratory.

Statistical analysis:

Sample size calculation:

Estimates were based on a two-tailed alpha test <0.05 and power of 80%. We sought to detect small to medium effect sizes between groups (stage and organ involvement) of equal size. With a sample size of 200 patients, with approximately 60-100 patients in each known group, we would have sufficient power to detect medium effect sizes of 0.3-0.5.

Data were summarized using means with SDs for continuous variables, frequencies with percentages for discrete variables, and mean scores with SD for PRO scores. Known clinical groups were compared by conducting analysis of variance to test for significant differences in mean scores across groups known to vary in disease severity. Cohen’s d was calculated for all pairwise comparisons. A Cohen’s d of <=0.2 was considered as a weak effect size, 0.3-0.5- medium, and >=0.8 as a large effect size.²⁴

Missing data:

22 values of NT-proBNP and 34 values of troponin were considered erroneous (input values “?” or “unk”, date instead of value). Because the questions in KCCQ-12 specifically list heart failure, we did not compare neurologic and renal clinical groups using the KCCQ-12. Only patients with cardiac involvement were provided the KCCQ-12 survey and thus the KCCQ-12 was ‘missing’ in 126 patients without reported cardiac AL. Analysis was conducted in R 4.0.3 software (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

The cohort characteristics are shown in Table 1. Median age at AL amyloidosis diagnosis was 60 years (range, 23-82) with a median time from diagnosis of 4.4 years (<0.1-21.9 years). Of these, 22% were within 2 years of diagnosis. Fifty eight percent of patients reported cardiac involvement and 39% reported involvement of 3 or more organs. Treatment included stem cell transplantation in 52% of individuals, and 50% of respondents were not on active therapy. Mean PRO scores and standard deviations are shown in Supplemental Table 1.

Table 1.

Baseline characteristics

Characteristic	N=297
Age at AL amyloidosis diagnosis, N = 292	60 (23.0-82.0)
Time from diagnosis to survey, years	4.4 (<0.1-21.9)
≤ 2	64 (22%)
>2	231 (79%)
Patient-reported race, N=294
White	264 (90%)
Black	10 (3.4%)
Other/Multiple	16 (5.4%)
Prefer not to answer	4 (1.4%)
Patient-reported ethnicity, N=295
Hispanic	9 (3.1%)
Non-Hispanic	278 (94%)
Prefer not to answer	8 (2.7%)
Sex, N=296
Male	141 (48%)
Female	155 (52%)
AL amyloidosis subtype, N=238
Lambda	165 (69%)
Kappa	71 (30%)
IgM	2 (0.8%)
Current 2004 AL stage, N=101
1	33 (33%)
2	37 (37%)
3	31 (31%)
Median (range) number of organs involved, N=293	2 (1-7)
Number of organs involved, N=293
1	126 (43%)
2	54 (18%)
3+	113 (39%)
Cardiac involvement	171 (58%)
Renal involvement	171 (58%)
Neurological involvement	90 (30%)
Hepatic involvement	34 (11%)
Gastrointestinal tract involvement	82 (28%)
Tongue involvement	47 (16%)
Skin/nail involvement	45 (15%)
Pulmonary involvement	13 (4.4%)
Muscle involvement	20 (6.7%)
Other organ involvement	28 (9.4%)
Treatment
Prior chemotherapy	262 (88%)
Prior stem cell transplant	153 (52%)
Currently on active treatment	146 (50%)

Comparison of PROs by clinical groups:

PRO comparison by Mayo 2004 stage grouping:

Stage at the time of survey completion was available in 101 patients where cardiac biomarker information was available. Table 2 shows the differences in mean PRO scores along with the Cohen’s d values for the three stage groups. The most discriminating PRO domains included physical function, fatigue, and social roles. For physical function, PROMIS and SF-36 subscales, but not KCCQ-12, showed similar significant discrimination by stage of disease with large effect sizes showing higher physical function with >3 point difference between stages I vs II vs III. For the domains of fatigue and social roles, PROMIS showed moderate and significant discrimination with >3 point difference with higher fatigue for stage III vs II vs I, but not SF-36 or KCCQ-12. PROMIS social roles also showed a greater than 3-point difference in scores and moderate effect size with higher social roles in stage I vs II vs III. The SF-36 GH subscale additionally also showed a small (<3 point) significant discrimination by stage with stage I vs II but not stage II vs III. No discrimination was seen for the domains of pain, sleep, or mental health. For summary scores, PROMIS GPHS and SF-36 PCS showed similar discrimination between the three stage groups, but not KCCQ-12.

Table 2.

Comparison of PROs by stage of disease using Mayo 2004 stage at the time of survey completion. Significant values are shown in bold

Domain	Measure	Stage I N=33	Stage II N=37	Stage III N=31	Cohen’s d III vs I	p- value*
Physical Function	PROMIS Physical Function-4	46.7	43.2	40.4	−0.76 (−1.26, −0.25)	0.01
	SF-36 Physical Functioning	49.3	46.5	42.2	−0.76 (−1.28, −0.25)	0.01
	SF-36 Role Physical	49.7	45.4	44.4	−0.52 (−1.03, 0.00)	0.10
	KCCQ-12 Physical Limitations	72.6	66.1	68.4	−0.19 (−0.87, 0.50)	0.7
Fatigue	PROMIS Fatigue-4	50.7	53.4	56.8	0.61 (0.11, 1.12)	0.05
	PROMIS Fatigue-8	51.2	53.4	57.5	0.64 (0.13, 1.14)	0.04
	SF-36 Vitality	48.8	46.4	44.4	−0.40 (−0.90, 0.11)	0.3
	KCCQ-12 Item 3	4.1	4.4	3.8	−0.16 (−0.84; 0.51)	0.5
Social Roles	PROMIS Ability to Perform Social Roles and Activities	51.0	47.2	44.9	−0.68 (−1.19, −0.18)	0.03
	SF-36 Social Function	49.7	47.7	44.6	−0.55 (−1.06, −0.04)	0.1
	KCCQ-12 Social Limitations	76.2	64.4	62.0	−0.54 (−1.22, 0.15)	0.3
Pain	PROMIS Pain Interference	47.7	51.5	51.7	0.42 (−0.08, 0.93)	0.2
	SF-36 Bodily Pain	54.7	50.4	52.2	−0.27 (−0.78, 0.23)	0.2
Sleep	PROMIS Sleep Disturbance	50.5	51.3	49.5	−0.10 (−0.60, 0.40)	0.8
Mental Health	PROMIS Anxiety	50.3	51.6	51.1	0.09 (−0.40, 0.59)	0.8
	PROMIS Depression	46.9	49.9	48.3	0.18 (−0.31, 0.68)	0.3
	PROMIS Cognitive Function	53.1	53.9	51.2	−0.25 (−0.76, 0.25)	0.3
	SF-36 Mental Health	54.9	53.4	53.4	−0.21 (−0.71, 0.30)	0.6
	SF-36 Role Emotional	52.5	51.4	48.4	−0.42 (−0.93, 0.09)	0.2
Summary scores and Other	PROMIS GPSS	47.4	43.2	41.6	−0.64 (−1.15, −0.13)	0.04
	PROMIS GMSS	51.8	47.3	46.9	−0.51 (−1.02, −0.01)	0.08
	SF-36 PCS	50.1	45.4	44.6	−0.62 (−1.14, −0.11)	0.03
	SF-36 MCS	51.8	49.2	48.2	−0.45 (−0.96, 0.06)	0.2
	SF-36 General Health	47.8	41.2	43.3	−0.45 (−0.96, 0.06)	0.03
	KCCQ-12 Total	73.3	67.2	64.6	−0.40 (−1.08, 0.28)	0.5
	KCCQ-12 Symptoms	73.1	72.2	68.3	−0.21 (−0.89, 0.47)	0.8
	KCCQ-12 Quality of Life	71.4	64.7	59.8	−0.40 (−1.08, 0.28)	0.5

^*One-way ANOVA

PROs by organ involvement:

For the type of organ involved, PROMIS and SF-36 were compared by presence of cardiac, renal, and neurological involvement (Table 3).

Table 3.

Comparison of PROs by amyloid organ involvement

Domain	Measure	Cardiac AL (Yes=171, N=126)		Neurologic AL (Yes=90, N=207)		Renal AL (Yes=171, No=126)
		Cohen’s d (yes vs no)	p- value	Cohen’s d (yes vs no)	p-value	Cohen’s d (yes vs no)	p- value
Physical Function	PROMIS Physical Function-4	−0.3	0.02	−0.3	0.02	0.1	0.6
	SF-36 Physical Functioning	−0.3	0.02	−0.2	0.1	0.1	0.5
	SF-36 Role Physical	−0.2	0.1	−0.3	0.03	−0.2	0.1
Fatigue	PROMIS Fatigue-4	0.2	0.1	0.3	0.02	0.2	0.2
	PROMIS Fatigue-8	0.1	0.3	0.2	0.08	0.2	0.1
	SF-36 Vitality	−0.3	0.01	−0.3	0.02	−0.1	0.4
Social Roles	PROMIS Ability to Perform Social Roles and Activities	−0.2	0.1	−0.3	0.03	−0.1	0.6
	SF-36 Social Function	−0.1	0.3	−0.2	0.1	−0.2	0.1
Pain	PROMIS Pain Interference	0.1	0.5	0.5	<0.001	0.3	0.01
	SF-36 Bodily Pain	−0.05	0.68	−0.51	<0.001	−0.3	0.03
Sleep	PROMIS Sleep Disturbance	0.2	0.1	0.3	0.02	0.05	0.7
Mental Health	PROMIS Anxiety	0	0.8	0	0.9	0.1	0.4
	PROMIS Depression	−0.1	0.5	0.2	0.1	0.1	0.3
	PROMIS Cognitive Function	0	>0.9	−0.2	0.3	−0.1	0.3
	SF-36 Mental Health	0	0.9	0	0.9	−0.3	0.04
	SF-36 Role Emotional	0.1	0.7	−0.1	0.5	−0.3	0.04
Summary scores and Other	PROMIS GPSS	−0.2	0.05	−0.4	<0.001	0	0.8
	PROMIS GMSS	−0.1	0.3	−0.3	0.03	−0.2	0.1
	SF-36 PCS	−0.3	0.03	−0.4	0.002	−0.1	0.3
	SF-36 MCS	−0.2	0.1	−0.3	0.06	−0.2	0.06
	SF-36 General Health	−0.3	0.01	−0.4	0.002	0.1	0.7

KCCQ-12 was only administered to individuals reporting cardiac amyloidosis and it is not relevant for renal or neurologic conditions. Therefore, it is not shown in this table

Interpretation of Cohen’s d: 0.2- small, 0.5- medium, and 0.8- large effect sizes [22]

Cardiac involvement:

Among patients with cardiac amyloid involvement compared to those without, for the domain of physical function, PROMIS and SF-36 showed significant discrimination with a small effect size. Whereas, for fatigue, only the SF-36 vitality subscale was significant with a medium effect size. The SF-36 GH subscale also showed small and significant discrimination by stage. For summary scores, both PROMIS GPSS and SF-36 PCS showed significant and small discrimination. Table 3 and Figure 1 shows the differences for the domains of physical function, social function, and fatigue by stage and by cardiac involvement.

Figure 1.

Comparison of fatigue, physical function, social roles, and physical health summary scores by stage and cardiac involvement

Neurologic involvement:

Comparisons of patients reporting neurologic involvement with those reporting no neurologic involvement showed significant discrimination for physical function by PROMIS physical function domain and SF-36 RP but not SF-36 PF subscale. For the domain of fatigue, SF-36 Vitality and PROMIS Fatigue-4 but not Fatigue-8 showed small and significant discrimination. PROMIS social roles but not SF-36 social function was discriminatory. For pain, both PROMIS and SF-36 showed significant discrimination with a moderate effect size. PROMIS sleep disturbance also showed a significant albeit small effect size. Finally, PROMIS GPSS and SF-36 PCS summary scores in addition to PROMIS GMSS and SF-36 GH also showed significant discrimination with a small effect size.

Renal involvement:

Comparisons of PROs by renal amyloid involvement showed no significant discrimination in physical function, fatigue, social roles, or summary scores, but there was small and statistically significant discrimination by pain for both PROMIS and SF-36. Additionally, both, SF-36 MH and RE subscales also showed a small and statistically significant discrimination in this group.

DISCUSSION

The utilization of PROs to enhance the quality of care for patients with AL amyloidosis is of great significance and value. However, while clinical researchers have several options to measure PROs in in AL amyloidosis, there is no clear guidance on which measure to choose. Additionally, standardization of PRO measurement across different settings is necessary. In this analysis, we compared three PRO measures, namely the KCCQ-12, PROMIS, and SF-36, to measure health-related quality of life in individuals with AL amyloidosis. Our findings demonstrate differences in clinically significant groups of disease involvement for these three measures in the same sample. Specifically, our results indicate that for the known clinical groups of stage, cardiac and neurologic organ involvement, physical function, fatigue, social roles, and the physical health summary scores show significant discrimination, with comparable effects between PROMIS and SF-36, but not the KCCQ-12.

Previous studies have demonstrated the reliability, construct validity, and sensitivity to change of the SF-36 in patients with AL amyloidosis.¹⁰ In addition, the KCCQ-12 has been qualified by the FDA for use in heart failure,¹⁴ while the NIH PROMIS measures enable the measurement of important domains of health across chronic diseases, including cancer.²⁵ The present study adds further support for the reliability and validity of the PROMIS and SF-36 measures as reasonable choices in the context of AL amyloidosis. In our previous work, which focused on a smaller cohort of newly diagnosed AL amyloidosis patients initiating active chemotherapy, PROMIS domains also shown evidence of responsiveness to change, further demonstrating construct validity, with changes in both hematologic response and NT-proBNP.¹² The current study, which is the first to compare the three measures in AL amyloidosis, indicates that the SF-36 and PROMIS measures perform comparably in this setting.

Conceptual models of disease manifestations and PROs in AL amyloidosis consistently identify many domains, including fatigue and social roles, as being impacted.^{5, 26} Other studies have also highlighted physical function as being severely impaired in this disease.^{6, 11} Therefore, while we present results for all domains, we were particularly interested in these three domains as concepts of interest for individuals living with AL amyloidosis. Our cross-sectional sample included heterogeneity in disease natural history, with individuals early in their disease course (i.e., within 2 years of diagnosis) and long-term (>5 years) survivors, individuals on active chemotherapy, and those who had a long treatment-free interval. Given that these patients continue to experience impaired health-related quality of life,²⁷ our approach remained relevant. Stage of disease is the strongest known predictor of outcomes in AL amyloidosis.³ When grouped by stage, PROMIS and SF-36 showed similar discrimination for physical function and summary scores, however PROMIS was better for the additional domains of fatigue and social roles. In terms of cardiac organ involvement, PROMIS and SF-36 were similar for physical function and physical health summary, but SF-36 performed better for fatigue, and neither measure showed discrimination for social roles. Among patients with neurologic amyloid involvement, in addition to physical function, fatigue, social roles, and physical summary score, pain, sleep disturbance, and mental summary scores are additionally discriminatory. Whereas, by renal amyloid involvement, physical function, social roles, fatigue, or summary scores were not discriminatory. Only pain by PROMIS and SF-36, and SF-36 MH and RE subscales show discrimination, albeit modest. This suggests that the pattern of amyloid involvement is an important factor to be considered when choosing PRO domains while designing clinical trials in this setting.

Our study has other limitations. We used the SF36v1 instead of v2 which has additional response choices for some items and improved wording. However, there is psychometric validity for both SF36v1 and SF-36v2 in AL amyloidosis.¹⁰ We were also limited by reliance of self-report of biomarkers. Though we allowed patients to upload laboratory results if available, we still had considerable missing data for cardiac biomarkers, and were thus unable to stage all patients. Furthermore, cardiac AL involvement can be a spectrum of changes which may or may not include heart failure. Organ involvement was also self-reported in this study and not confirmed with health records. Thus, the weak performance of the KCCQ-12 in this study should be interpreted with caution.

In summary, our cross-sectional community-based study provides a comparison of three PRO measures by clinical groups in AL amyloidosis. PROMIS and SF-36 measures demonstrate acceptable effect sizes to discriminate between clinical groups in AL amyloidosis patients. Our findings suggest that the domains of physical function, fatigue, social roles, and the physical health summary scores are particularly useful for discriminating between stage, cardiac and neurologic involvement, but not renal involvement. However, it is important to note that patients with AL amyloidosis can experience many other symptoms such as dyspnea, edema, and dizziness), which should also be considered in PRO measurement in this context of use. Additionally, an assessment of disease-specific symptoms along with the generic domain measures, may be helpful in providing a more comprehensive understanding of health-related quality of life in AL amyloidosis patients.

Data availability statement:

Data will be shared upon reasonable request to the corresponding author