The severity of systemic amyloidosis is thought to be related to the extent of amyloid deposition. We studied whether amyloid load in fat tissue reflects disease severity and predicts survival.
The severity of systemic amyloidosis is thought to be related to the extent of amyloid deposition. We studied whether amyloid load in fat tissue reflects disease severity and predicts survival.
We studied all consecutive patients with systemic amyloidosis seen between January 1994 and January 2007 in our tertiary referral center. Congo red–stained abdominal fat smears were graded by 2 observers using a validated semiquantitative scoring system. Disease severity was measured by the total number of major organs involved and the extravascular retention of the serum amyloid P component (EVR24). The association of amyloid load in fat tissue with disease severity and overall survival was studied using multiple regression analysis.
Two hundred twenty patients were included in the study (120 with AL amyloidosis, 66 with AA amyloidosis, and 34 with ATTR amyloidosis). Amyloid grade in fat tissue was associated with the number of major organs involved and EVR24. Female sex turned out to be associated with a higher grade of amyloid in fat tissue than male sex. Amyloid grade in fat tissue was an independent predictor of decreased survival, as were heart involvement, the number of organs involved, AA or AL type of amyloid, and age.
The amount of amyloid in subcutaneous fat tissue in systemic amyloidosis reflects disease severity, as measured by the number of organs involved and EVR24, and predicts decreased survival independent of other well-known factors.
The amyloidoses are a group of diseases characterized by deposition of protein fibrils with a cross–β-pleated sheet molecular structure. This structure accounts for the binding affinity of Congo red dye and the green birefringence in polarized light (1). The deposition of amyloid fibrils results in loss of organ function. In systemic amyloidosis, major organ systems such as the heart, kidneys, liver, and nervous system may be involved. Therefore, systemic amyloidosis is a disease with high morbidity and mortality.
The diagnosis of amyloidosis is based on histologic analysis, demonstrating positive staining with Congo red and characteristic apple-green birefringence. Abdominal fat tissue aspiration for the detection of amyloid is a safe and simple method with good diagnostic accuracy, i.e., specificity of 100% and sensitivity up to 93% (2). The 3 major types of systemic amyloidosis are AA, AL, and ATTR amyloidosis. These types differ with regard to the nature of the precursor proteins: serum amyloid A protein in AA amyloidosis, immunoglobulin free light chain in AL amyloidosis, and transthyretin in ATTR amyloidosis. Because of the different nature of the precursor protein for each type of systemic amyloidosis, the underlying diseases require different treatment (3).
Opportunities for treatment and prognosis highly depend on the severity of organ involvement at the time of diagnosis. Disease severity can be assessed by the number of major organs involved and by 24-hour tissue retention of the 123I-labeled serum amyloid P (SAP) component (4). These prognostic markers are especially useful for defining risk groups in relation to therapy options. Recently, a semiquantitative grading system of amyloid deposition in fat tissue has been validated in patients with AA amyloidosis (5). However, no information is available on whether the amount of amyloid deposition in fat tissue correlates with the total load of amyloid deposition throughout the body, and therefore might have prognostic value.
In the present prospective study, we analyzed semiquantitatively assessed amyloid deposition in subcutaneous fat tissue and studied whether it was related to disease severity and predictive for overall survival in patients with systemic amyloidosis.
A prospective observational study was conducted in all of the consecutive patients with systemic amyloidosis who were seen between January 1994 and January 2007 at our tertiary referral center, and who were followed until January 2009. The local Ethics Committee approved the study and informed consent was obtained from all of the patients. Systemic amyloidosis was diagnosed either by the detection of amyloid in a biopsy site typically involved in systemic amyloidosis, such as the abdominal fat tissue, kidney, liver, spleen, or nerve, or by positive biopsy results derived from at least 2 different sites, including the heart, gastrointestinal tract, and bone marrow.
AA, AL, or ATTR amyloidosis was classified as follows: AA amyloidosis was distinguished immunohistochemically, using monoclonal murine anti-human AA antibodies. AL amyloidosis was defined by detecting a clonal plasma cell dyscrasia in patients immunohistochemically negative for AA amyloidosis. A clonal plasma cell dyscrasia was diagnosed when a κ or λ free light chain was detected by immunofixation electrophoresis or free light chain assay in serum and/or urine, or when a relative excess of plasma cells producing 1 of the 2 light chains (κ or λ) was detected in the bone marrow. In patients with only cardiomyopathy or neuropathy, a mutation in the transthyretin gene was excluded before the diagnosis of AL amyloidosis was made. ATTR amyloidosis was defined by detection of a transthyretin mutation in patients whose amyloid deposits stained specifically with anti-transthyretin antibodies (6).
Subcutaneous abdominal fat tissue was aspirated and stained with Congo red dye as previously described (2, 7). In all of the patients, at least 30 mg of fat tissue was aspired. If a patient had more than one fat tissue biopsy, the first sample was used. All of the fat smears were reviewed by 2 experienced observers (BPCH, JB). They independently scored all of the smears at 40× magnification, 3 smears per patient with the total screening time being 5 minutes, using a 100W Olympus BX50 microscope (Olympus, Hamburg, Germany). During this procedure, the polarization filter was turned almost continuously in order to detect even minimal amyloid deposits. A validated semiquantitative grading system was used for each patient, ranging from 0 to 4+ as shown in Figure 1: 0 = negative; 1+ = minute, <1% of the surface area; 2+ = little, between 1% and 10%; 3+ = moderate, between 10% and 60%; and 4+ = abundant, >60% (5). In the case of disagreement between the 2 observers, the 3 smears were reviewed and discussed to obtain consensus. All of the fat smears were assessed for clinical data in a blinded manner. The interobserver agreement was 0.74 (κ = 0.61), the intraobserver agreement was 0.75 (κ = 0.67), and no difference of more than 1 grade was observed.
At the time of the semiquantitative assessment of amyloid in subcutaneous fat tissue at diagnosis, organ involvement was assessed using established criteria with minor modifications (8). Organ involvement was assessed in 4 major categories, as in internationally excepted consensus: heart, kidney, liver, and nerve involvement. Heart involvement was defined as the presence of heart failure (New York Heart Association grade >2), low voltage on electrocardiography, or a mean left ventricular wall thickness >12 mm in the absence of a history of hypertension or other potential causes of left ventricular hypertrophy. Kidney involvement was defined as renal function disturbance (endogenous creatinine clearance <60 ml/minute) or proteinuria >0.5 gram/day, without overt other explanation. Liver involvement was defined as hepatomegaly (physical examination confirmed by ultrasound span >16 cm) or an increase in alkaline phosphatase level >180 units/liter (upper limit of normal 120). Nerve involvement was defined as abnormal autonomic nervous function testing using validated tests based on Ewing and Clark or abnormal sensory or motor findings on neurologic examination (9).
In all of the patients, SAP scintigraphy was performed at the time of the semiquantitative assessment of amyloid in fat tissue (6). Tissue retention of 123I-labeled SAP after 24 hours (EVR24) was measured in all of the patients as a mark of the amyloid load of the body (4). Because the only excretion route for SAP is urinary excretion, the whole-body retention of SAP after 24 hours was calculated by subtracting the cumulative activity excreted in the urine as a percentage of the injected dose and subtracting this dose from 100%. Total body extravascular (or tissue) retention was determined at 24 hours after injection by subtracting plasma activity from whole-body retention, both expressed as the percent injected dose. EVR24 reference control values were <50% (4). In all of the patients, EVR24 was measured as an indicator of the amyloid load of the body (10). One must realize that myocardium does not show specific uptake of SAP, perhaps because of the combination of high background activity of the tracer still present in the blood pool and decreased permeability in the cardiac tissue of this tracer with a high molecular weight.
SPSS statistical software, version 16.0.1 (SPSS, Chicago, IL), was used. Frequencies were compared with the use of chi-square tests or Fisher's exact test, differences between groups were tested with the Mann-Whitney test, and correlation was tested with Spearman's rank test. Analysis of variance (ANOVA) was used for the groups, where appropriate. Overall survival for the different amyloid grades of fat tissue was studied with Kaplan-Meier curves and evaluated with the log rank test. In all of the tests, 2-tailed P values less than 0.05 were considered significant. Amyloid grade of fat tissue, sex, age, amyloid type, amyloid duration, and EVR24 were studied with the number of organs as the dependent variable using ordinal regression, and the pseudo-R2 (Nagelkerke) was calculated (11). The same variables (with the number of organs instead of EVR24) were studied with EVR24 as a dependent variable using linear regression. Predictors of survival were identified in Cox proportional hazards analysis using survival as a dependent variable using the same variables as mentioned above, extended with the 4 specific organs. Multiple regression was performed with a stepwise forward regression model, with an entry probability for each variable set at 0.05. Two models were used for multiple regression in the Cox analysis because the number of major organs involved and specific organ involvement are related topics. In one model, heart involvement was evaluated, and in the other model, heart involvement was replaced by the number of organs.
A total of 220 patients with systemic amyloidosis were included in the study: 66 with AA amyloidosis, 120 with AL amyloidosis, and 34 with ATTR amyloidosis. Patient characteristics are listed in Table 1. In these 220 patients, kidney involvement was the most prominent feature (75%), as seen in the majority of patients with AL (82%) and AA (89%) amyloidosis. The heart was more frequently involved in AL and ATTR amyloidosis (P < 0.001) compared with AA amyloidosis. In AL amyloidosis, this was mostly reflected by clinical heart failure and in ATTR amyloidosis by left ventricular wall thickness. As expected, neuropathy was most prominent in ATTR amyloidosis (85%). In patients with AL amyloidosis, the liver was more frequently involved and the number of organs was higher compared with AA and ATTR amyloidosis (P < 0.001). Patients with AL amyloidosis were older than patients with ATTR amyloidosis (P < 0.001). Neither the duration of tissue-proven amyloidosis nor amyloid grade in fat tissue differed among the 3 subgroups. Higher grades of amyloid in fat tissue were found in women than in men (P < 0.0001), as shown in Figure 2.
|AL amyloidosis (n = 120)||AA amyloidosis (n = 66)||ATTR amyloidosis (n = 34)||All types (n = 220)|
|BMI, kg/m2||24 (15–36)||23 (14–40)||23 (17–29)||23 (14–40)|
|Age, years||62 (33–84)||56 (13–85)||53 (24–77)||58 (13–85)|
|Months of amyloidosis||2 (0–167)||2 (0–467)||3 (0–140)||2 (0–467)|
|Months of followup||44 (2–142)||71 (0–141)||72 (1–152)||55 (0–152)|
|Involved major organs, no. (%)|
|Heart||66 (55)||11 (17)||13 (38)||90 (41)|
|Kidney||98 (82)||59 (89)||8 (23)||165 (75)|
|Liver||37 (31)||5 (7)||0 (0)||42 (19)|
|Nerve||53 (44)||11 (17)||29 (85)||93 (42)|
|Organs per patient, no. (%)|
|0||3 (2)||4 (6)||2 (6)||9 (4)|
|1||31 (26)||44 (67)||16 (47)||91 (41)|
|2||44 (37)||12 (18)||14 (41)||70 (32)|
|3||33 (28)||6 (9)||2 (6)||41 (19)|
|4||9 (7)||0 (0)||0 (0)||9 (4)|
|Amyloid grade in fat tissue, no. (%)|
|0||7 (6)||3 (4)||3 (9)||13 (6)|
|1+||15 (12)||13 (20)||4 (12)||32 (15)|
|2+||18 (15)||10 (15)||4 (12)||32 (15)|
|3+||37 (31)||23 (35)||11 (32)||71 (32)|
|4+||43 (36)||17 (26)||12 (35)||72 (33)|
The number of organs correlated with amyloid grade in fat tissue (rs = 0.25, P < 0.01), as illustrated in Figure 2. The outcome of multiple ordinal regression with the number of organs as the dependent variable was that age (P < 0.05), amyloid grade in fat tissue (P < 0.05), AL type of amyloid (P < 0.001), and EVR24 (P < 0.001) were included in the model, yielding a pseudo-R2 of 0.37 (Nagelkerke) (11). Sex, body mass index (BMI), and amyloid duration were excluded from the model.
The ANOVA test (P = 0.01), followed by a posttest for linear trend (P < 0.005), showed a linear trend of the EVR24 for the amyloid grades in fat tissue, as shown in Figure 3. The median percentage EVR24 was 37, 42, 50, 54, and 56 for amyloid grades 0, 1+, 2+, 3+, and 4+, respectively. The outcome of multiple linear regression with EVR24 as the dependent variable was that amyloid grade in fat tissue (P < 0.05), BMI (P < 0.05), AL or AA type of amyloid (P < 0.001), and number of organs (P < 0.001) were included in the model, yielding an adjusted R2 of 0.30. Sex, age, and amyloid duration were excluded from the model.
Survival differed among the groups with different amyloid grades in fat tissue (P < 0.001), as shown in Figure 4 by the Kaplan-Meier curves for 3 equally sized groups (0–2+, 3+, and 4+ with median survival of >60, 31, and 19 months, respectively). Multiple regression Cox proportional hazards analysis demonstrated that amyloid grade in fat tissue is a predictor for survival. Other predictors for survival were heart involvement (model 1) or the number of major organs (model 2), AA or AL type of amyloidosis, and age. Details are shown in Table 2.
|HR (95% CI)||P|
|Amyloid grade (CR range 0–4) of fat tissue||1.34 (1.15–1.57)||< 0.001|
|Amyloid duration, months||1.00 (0.99–1.00)||0.160|
|Male sex||1.20 (0.85–1.69)||0.296|
|Age, years||1.03 (1.02–1.05)||< 0.001|
|BMI, kg/m2||0.97 (0.92–1.02)||0.176|
|Type of amyloid|
|AL or ATTR versus AA||1.02 (0.71–1.47)||0.904|
|AL or AA versus ATTR||2.34 (1.36–4.03)||0.002|
|AL versus AA or ATTR||1.64 (1.16–2.32)||0.005|
|Heart||3.08 (2.17–4.37)||< 0.001|
|Number of organs involved (range 0–4)||1.65 (1.39–1.96)||< 0.001|
|EVR24, %||1.02 (1.01–1.02)||< 0.001|
|Amyloid grade (CR range 0–4) of fat tissue||1.33 (1.13–1.57)||0.001|
|Age, years||1.02 (1.00–1.04)||0.028|
|AL or AA type of amyloid||2.48 (1.40–4.39)||0.002|
|Heart involvement||2.88 (2.01–4.12)||< 0.001|
|Amyloid grade (CR range 0–4) of fat tissue||1.36 (1.16–1.60)||< 0.001|
|Age, years||1.02 (1.00–1.04)||0.017|
|AL or AA type of amyloid||1.90 (1.08–3.36)||0.026|
|Number of organs involved (range 0–4)||1.55 (1.30–1.85)||< 0.001|
The present study shows a significant association between the semiquantitatively assessed amyloid load in the subcutaneous fat tissue of patients with systemic amyloidosis and disease severity, as defined by the number of organs involved and EVR24. Higher amyloid grade in fat tissue is associated with more severe disease, independent of the type of amyloid and age. Moreover, in systemic amyloidosis, the amyloid load in fat tissue is a predictor for overall survival, also independent of the type of amyloid and age. Therefore, semiquantitative grading of amyloid in Congo red–stained fat tissue reflects the severity of amyloid deposition throughout the body and may be a useful additional method for estimating risks and prognosis in systemic amyloidosis.
Heart involvement is still the major predictor of survival, as known from the literature (12, 13). The assessment of heart involvement in systemic amyloidosis is sometimes difficult, but recent markers such as N-terminal pro–brain natriuretic peptide (NT-proBNP) and cardiac troponin have enriched the diagnostic arsenal. Most studies concerning NT-proBNP and cardiac troponin in amyloidosis have been performed in the AL type of amyloidosis (12, 13). A recently published study showed that cardiac troponin is not useful for detecting heart involvement in ATTR amyloidosis (14). Because these biomarkers were not available during our study period, it would be very interesting to correlate the cardiac biomarkers with amyloid load in fat tissue in a future study concerning risks and prognosis in systemic amyloidosis. An alternative for heart involvement is the number of organs, and this variable is also an important predictor of survival. Therefore, assessing heart involvement and the number of organs remain important factors in estimating prognosis in systemic amyloidosis.
An interesting finding is that women have a significantly higher amount of amyloid in subcutaneous fat tissue than men. This sex-specific distribution of amyloid in fat tissue might be related to sex hormones, but the exact mechanism is unknown. The BMI did not explain the sex-specific difference. Although women have a higher grade of amyloid in fat tissue, we did not find a difference in survival between men and women in any of the 3 types of systemic amyloidosis. Women seem to deposit amyloid more specifically in fat tissue. Recently, Rapezzi et al described that in ATTR amyloidosis with heart involvement, women tended to have less severe heart disease than men (15). Therefore, sex-specific differences in amyloid deposition may be present among different tissues.
Limitations of our study are, first, the use of a semiquantitative instead of a quantitative fat tissue analysis method, although this semiquantitative analysis was validated in a previous study (5). Second, amyloid load in vital organs was determined by clinical organ involvement and not by biopsy, but the latter would have been unrealistic and risky. A completely different and noninvasive alternative would have been SAP scintigraphy to obtain an overall view of the amyloid load of the body (6). Limitations of this approach, however, are a lack of visible uptake of SAP in major organ systems such as the heart and nerve tissue. Therefore, we studied the extravascular retention of SAP (EVR24), but this method is less precise because of the wide range of values among individual patients. Third, amyloid load in fat tissue as a prognostic factor for individual patients and for defining risk groups in relation to therapy needs further study.
Since this study confirms that the amount of amyloid in fat tissue reflects the severity of amyloidosis, future study objectives are to establish the regression or disappearance of amyloid in fat tissue after successful treatment of systemic amyloidosis and the rate of this process (16, 17). In a recently published small landmark study, we were able to show regression of amyloid in fat tissue after successful treatment of AL amyloidosis. This turned out to be a relatively slow process, starting as of 10 months following treatment, with significant regression of amyloid in fat tissue in 50% of patients after 2.4 years and in 80% of patients after 3.2 years (17). Improvement of organ function in systemic amyloidosis might take months to years, especially when the heart is involved. It would be interesting to know whether regression of amyloid in fat tissue could be used as a measure of effect of therapy and a predictor of further organ improvement. The difference of amyloid load in fat tissue between men and women is intriguing. Finding the cause of this sex-specific difference might reveal permissive or rate-limiting steps in amyloid deposition in the extracellular matrix, not only in fat tissue but also in essential tissues.
In conclusion, semiquantitative assessment of Congo red–stained fat tissue is of prognostic value in systemic amyloidosis, irrespective of the type of amyloidosis. Higher amyloid amount in fat tissue reflects more severe disease and predicts decreased survival independent of other well-known factors.
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. van Gameren 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. Van Gameren, Hazenberg, Bijzet, Haagsma, Vellenga, Posthumus, Jager, van Rijswijk.
Acquisition of data. Van Gameren, Hazenberg, Bijzet, Haagsma, Vellenga, Posthumus, Jager, van Rijswijk.
Analysis and interpretation of data. Van Gameren, Hazenberg, Haagsma, Vellenga, Posthumus, Jager, van Rijswijk.