Nephrol Dial Transplant (2016) 31: 275–283
doi: 10.1093/ndt/gfv261
Advance Access publication 30 June 2015
Renal impairment in patients with polyneuropathy,
organomegaly, endocrinopathy, monoclonal gammopathy and
skin changes syndrome: incidence, treatment and outcome
Wei Ye1,*, Chen Wang2,*, Qian-Qian Cai2, Hao Cai2, Ming-Hui Duan2, Hang Li1, Xin-Xin Cao2,
Dao-Bin Zhou2 and Jian Li2
1
Department of Nephrology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College,
Beijing, China and 2Department of Hematology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union
Medical College, Beijing, China
Correspondence and offprint requests to: Jian Li; E-mail: [email protected]
*
W.Y. and C.W. contributed equally to this work.
Background. Polyneuropathy, organomegaly, endocrinopathy,
monoclonal gammopathy and skin changes (POEMS) syndrome is a multisystem disorder arising from underlying plasma cell dyscrasia. Renal impairment and related pathological
changes have been reported, but data on its prevalence, response to therapy and impact on survival are still lacking.
Methods. We retrospectively reviewed 299 patients diagnosed
with POEMS syndrome in a tertiary-care university hospital
from 2000 until 2014. The estimated glomerular filtration rate
(eGFR) was used to define renal impairment and response,
according to International Myeloma Working Group criteria.
We examined the impact of renal impairment and response
on patient survival.
Results. Sixty-seven patients (22.4%) had renal impairment
(eGFR < 60 mL/min/1.73 m2) at baseline. In a multivariate
analysis, ascites was independently associated with renal impairment [odds ratio (OR) 12.366, P < 0.001]. Renal impairment was reversible in 66.0% of patients receiving therapy
and was associated with a shorter time interval between symptom onset and treatment (OR 0.059, P = 0.043) and a vascular
endothelial growth factor remission (OR 15.958, P = 0.050) in a
multivariate analysis. In terms of therapy, patients with a renal
response more commonly received a novel agent-based regimen (P = 0.037), which also led to a shorter response time
(P = 0.001). With a median follow-up of 27.4 months, inferior
survival was observed in patients with severe renal impairment
(eGFR < 30 mL/min/1.73 m2), but not in those with moderate dysfunction (eGFR 30–59 mL/min/1.73 m2), compared
© The Author 2015. Published by Oxford University Press
on behalf of ERA-EDTA. All rights reserved.
with patients without renal impairment. A renal response, if
achieved, predicted improved survival.
Conclusions. Renal impairment is a common complication of
POEMS syndrome, but can be reversed with effective therapy in
most cases.
Keywords: POEMS syndrome, renal impairment, vascular
endothelial growth factor
INTRODUCTION
Polyneuropathy, organomegaly, endocrinopathy, monoclonal
gammopathy and skin changes (POEMS) syndrome is a multisystem disorder encountered across various medical disciplines
[1, 2]. Although not included in the acronym, renal manifestations have been reported, presenting as proteinuria, hematuria
and renal dysfunction. Various underlying glomerular changes,
such as membranoproliferative glomerulonephritis-like lesions,
microangiopathic lesions and mesangiolytic lesions, have been
described [3, 4]. Increased levels of vascular endothelial growth
factor (VEGF), a potent angiogenic factor with a presumed
pathogenic role in POEMS syndrome, may also contribute to
these renal lesions [5, 6].
However, these findings are mainly derived from sporadic
reports, and a more general picture of the renal impairment associated with POEMS syndrome is still lacking, especially its
impact on the disease course and response to therapy. We
sought to define the prevalence, clinical correlations, treatment
responses and prognostic significance of renal impairment in a
retrospective cohort of patients with POEMS syndrome. The
275
ORIGINAL ARTICLE
A B S T R AC T
pathogenic role of VEGF in renal impairment was also
addressed in this study.
ORIGINAL ARTICLE
M AT E R I A L S A N D M E T H O D S
Patients
A comprehensive search of our medical records was performed to identify patients with a diagnosis of POEMS syndrome at Peking Union Medical College Hospital between
January 2000 and November 2014. Two of the authors (C.W.
and J.L.) reviewed the data, and 299 patients were included in
this study, who met the diagnostic criteria proposed by Dispenzieri [2]: (i) the presence of both polyneuropathy and monoclonal gammopathy; (ii) the presence of one of the following three
major criteria: Castleman’s disease, osteosclerosis or elevated
VEGF and (iii) one of the six minor criteria: organomegaly,
extravascular volume overload, endocrinopathy, skin changes,
papilledema and either thrombocytosis or polycythemia.
The detailed clinical features of POEMS syndrome were recorded at diagnosis. The overall neuropathy limitation scale
(ONLS) was used to assess neurological disability, as described
previously [7]. Nerve conduction studies and electromyography
(NCS/EMG) were not performed regularly, but were prescribed
for patients without typical neurological findings, e.g. who scored
0 in the ONLS assessment. Laboratory data associated with
POEMS syndrome and renal impairment were collected, including those from a bone-marrow examination, serum protein electrophoresis, serum and urine immunofixation, serum creatinine,
urinalysis and 24 h urine protein (from a 24 h urine collection).
The estimated glomerular filtration rate (eGFR) was calculated
using the Chronic Kidney Disease Epidemiology Collaboration
(CKD-EPI) equation [8]. Serum VEGF was measured with a
human Quantikine ELISA Kit (R&D Systems, Minneapolis,
MN, USA), and serum VEGF < 600 pg/mL was considered normal, as described elsewhere [7, 9].
The treatment and follow-up data were retrieved from medical records. The authors also tried to contact all the patients
and treating physicians to obtain follow-up data, including
data on deaths and causes of death, via telephone, mail or
e-mail. The therapeutic regimens included a melphalan-based
regimen (69 patients), autologous stem cell transplantation (88
patients), a novel agent-based regimen (lenalidomide in 95 patients, thalidomide in 3 patients and bortezomib in 3 patients,
in total 101 patients) and others (miscellaneous treatments such
as prednisone or dexamethasone alone, traditional Chinese
medicine or no treatment) (41 patients). Besides antiplasma
cell therapy, supportive treatments were given when indicated,
including intravenous hydration, alkalization of the urine and
discontinuation of known nephrotoxic agents. No patient required dialysis. The treatment responses were evaluated in
those patients with at least one follow-up consultation
and the corresponding laboratory tests at our hospital. An
expert-recommended test schedule was used [2], but its implementation was also adjusted on the basis of patient compliance.
Briefly, hematological parameters, including complete blood
count, blood chemistry (e.g. creatinine) and M protein (e.g.
serum protein electrophoresis and serum immunofixation
276
electrophoresis), were measured every 1–2 weeks when the patient was hospitalized and at every follow-up visit. Serum levels
of VEGF were measured every 1–3 months in the early course
of treatment and every 3–6 months in the further follow-up.
All patients gave their informed consent, and the study was
approved by the Institutional Review Board of Peking Union
Medical College Hospital, in accordance with the Declaration
of Helsinki.
Response assessment
Renal response. Renal impairment was defined as an eGFR
<60 mL/min/1.73 m2 using the CKD-EPI formula, based on
the definition of at least moderate renal dysfunction by the
National Kidney Foundation Kidney Disease Outcomes Quality
Initiative [10]. This definition was applicable to patients with
stabilized serum creatinine levels, despite volume replacement
if there was volume insufficiency, before the initiation of effective anti-plasma cell therapy. The stabilized creatinine value was
used to calculate eGFR. The degree of renal response was evaluated according to eGFR change after treatment. Briefly, a complete response (CR) was defined as an increase in baseline eGFR
to >60 mL/min/1.73 m2. A partial response (PR) was defined as
an increase in eGFR from <15 to 30–59 mL/min/1.73 m2.
Because such improvements are clinically relevant and all but
one patient achieved CR, we focussed our analysis on these
two responses (CR and PR) and collectively defined them
as ‘a renal response’ in general. Definitions of both renal impairment and renal response were borrowed from studies
of myeloma-associated renal impairment, according to the
consensus statement on behalf of the International Myeloma
Working Group, which were further updated in their recent
studies [11–13]. In contrast, renal response criteria developed
for light chain amyloidosis were not used, as at least 0.5 g reduction of 24-h urine protein was required to fulfill the response
and massive proteinuria was not common in our patients.
Responses of other systems. A complete hematological response was defined as the complete disappearance of the monoclonal protein in both serum and urine specimens [7]. A CR for
serum VEGF was the normalization of VEGF (<600 pg/mL)
[14]. No other levels of response were specified.
Renal histopathology
Five patients in this retrospective cohort underwent renal
biopsy at our hospital. The renal histopathology data, including glomerular, tubulointerstitial and vascular changes, were
evaluated by two nephropathologists (W.Y. and H.L.) without
access to the clinical information. The results of immunofluorescent and Congo red staining were also described. All renal
biopsy specimens contained more than 10 glomeruli.
Statistical analysis
Data analysis was performed with the statistical software
package SPSS 22.0 (SPSS, Inc., Chicago, IL, USA). Differences
between various groups were compared with the χ2 test for categorical variables (using Fisher’s exact test when appropriate)
and with the Mann–Whitney test for continuous variables. A logistic regression analysis was used for the multivariate analysis,
W. Ye et al.
including all variables that were significant (P < 0.10) in the univariate analysis. Time to response was calculated from the date of
initiation of treatment until the date when the criteria for response were first met. Patients who died before a response
could be evaluated were censored at the time of their death. Survival was evaluated from the date of treatment until the date of
death or the last follow-up. Survival curves were plotted with the
Kaplan–Meier method and compared with a log-rank test.
A landmark analysis at 3 months was performed to evaluate patients who survived long enough to benefit from treatment.
Propensity-score 1 : 1 matching was used to eliminate any imbalance in the variables between groups, using calipers equal to
0.3 of the standard deviation of the estimated propensity score
without replacement. We matched the age and sex of the patients
with and without renal impairment, to identify the disease features associated with renal impairment at baseline. Comparisons
between groups were made both before and after matching. All
data were considered statistically significant at P ≤ 0.05.
R E S U LT S
Pathological features
The pathological changes in the patients are described in
Table 1. The most frequently observed glomerular changes
were mesangial and endothelial cell proliferation. The glomerular endothelial cells were swollen and occluded the capillary
lumens (Figure 1A ). Additional findings included thickening
of the glomerular basement membrane and double contour
formation. In the tubulointerstitium, we noted interstitial fibrosis with focally scattered infiltration of mononuclear cells
and tubular atrophy. Hyalinosis of the arterioles, arteriolar
wall thickening and luminal occlusion were occasionally observed. Immunofluorescent staining showed no deposition of
immunoglobulin or complement. Congo red staining was also
negative in the three cases examined. A common finding under
electron microscopy was a widening of the subendothelial
space, but there were no electron-dense deposits (Figure 1B).
Table 1. Clinicopathological summary of POEMS patients with renal biopsy
Patient 1
Patient 2
Patient 3
Patient 4
Patient 5
Clinical findings
Sex/age (years)
Disease onset to biopsy (months)
Previous therapy
Creatinine (μM)
eGFR (mL/min/1.73 m2)
24-h urine protein (g)
Microhematuria
Ascites
Serum VEGF (pg/mL)
M/41
4
—
132
66
3.0
−
+
5640
M/66
12
Prednisone
117
64
0.5
+
−
NA
F/44
14
—
108
62
0.2
+
+
NA
F/63
16
—
110
53
2.5
−
+
NA
M/62
7
Prednisone
105
76
0.3
−
+
3042
Renal pathology
Glomeruli
Number
Cellularity
Mesangial proliferation
Endothelial proliferation
Double contour
GBM thickening
Tubulointerstitial changes
Vessel changes
31
+++
+++
++
+++
+++
++
−
22
+++
+++
+++
++
Focal
+
Hyalinosis
19
+++
+++
++
+++
++
+++
−
19
++
++
±
−
Focal
++++
−
−
NA
−
−
−
−
31
++
+++
+++
+++
+++
++++
Arteriolar wall thickening
and luminal occlusion
−
−
Immunofluorescent staining
Congo red staining
−
NA
VEGF, vascular endothelial growth factor; GBM, glomerular basement membrane; NA, not available.
Renal impairment in POEMS syndrome
277
ORIGINAL ARTICLE
Clinical characteristics
The patient characteristics are shown in Supplementary data,
Table S1. One hundred ninety-one patients (63.9%) were male.
The median age at diagnosis was 47 years (range, 21–74). Polyneuropathy (ONLS score: median 4, range, 0–11) and monoclonal gammopathy (IgA heavy chain isotype in 65.2% cases and λ
light chain isotype in all cases) were seen in all patients. In 16 patients (5.4%) with an ONLS score of 0, NCS/EMG was performed, and slowing of nerve conduction was observed. In
terms of other major disease features, 156 had osteosclerosis
and 43 of 73 who had a lymph-node biopsy had Castleman’s disease. Serum VEGF levels were measured in 235 patients, and the
median value was 3894 pg/mL (range, 111–22 993 pg/mL). Two
hundred twenty-one patients (94.0%) had serum VEGF levels
>600 pg/mL. Other relevant features were observed in various
proportions of patients, as shown in Supplementary data,
Table S1. With respect to renal manifestations, microhematuria
was documented in 29 patients (9.7%), and the median value for
24-h urine protein was 0.3 g (range, 0–3.0 g; >1.0 g, n = 17,
5.7%). Sixty-seven patients (22.4%) had baseline eGFR < 60
mL/min/1.73 m2: in 52, eGFR was 30–59 mL/min/1.73 m2; in
14, eGFR was 15–29 mL/min/1.73 m2; and in only one patient,
baseline eGFR was <15 mL/min/1.73 m2 in the kidney-failure
range.
ORIGINAL ARTICLE
Comparison of patients with and without renal
impairment
Table 2 shows the features of patients with and without renal
impairment at baseline. Patients with renal impairment, compared
with those without renal impairment, were older (age >50 years,
58.2 versus 37.9%, P = 0.004) and were less frequently male
(49.3 versus 68.1%, P = 0.005). Other POEMS features, including hepatomegaly (P = 0.045), hypothyroidism (P = 0.018),
ascites (P < 0.001), pulmonary hypertension (P = 0.013) and
osteosclerosis (P = 0.019), also significantly differentiated
in the univariate analysis. Microhematuria (19.4 versus 6.9%,
P = 0.003), but not proteinuria (24 h urine protein >1.0 g, 10.4 versus 4.3%, P = 0.056), was statistically more common in the renal
impairment group. When we adjusted the differences between
the groups in the multivariate analysis, age >50 years [odds ratio
(OR) 4.310, 95% confidence interval (CI) 2.093–8.878, P < 0.001],
ascites (OR 12.366, 95% CI 5.046–30.306, P < 0.001) and
F I G U R E 1 : Renal biopsy at diagnosis. (A) Representative light-microscopic image of glomeruli with periodic acid–Schiff–methenamine silver
staining (original magnification, ×400). Endothelial cell proliferation (red arrowhead) and mesangial matrix expansion with double contour
formation (red arrow) are labeled. (B) Electron micrograph. Expansion of the subendothelial space (white arrowhead) and thickening of the lamina
rara interna (white arrow) are noted.
Table 2. Characteristics of patients with and without renal impairment at baseline
RI (n = 67)
Non-RI (n = 232)
P-value*
RI (n = 65)
Matched non-RI
(n = 65)
P-value*
Demographic information
Age >50 years (N, %)
Male (N, %)
39 (58.2)
33 (49.3)
88 (37.9)
158 (68.1)
0.004
0.005
37 (56.9)
32 (49.2)
38 (58.5)
31 (47.7)
0.859
0.861
POEMS features
ONLS ≥ 4 (N, %)
Hepatomegaly (N, %)
Splenomegaly (N, %)
Lymphadenopathy (N, %)
Castleman’s disease (N, %)a
Hypothyroidism (N, %)
Diabetes (N, %)
Hyperpigmentation (N, %)
Angioma (N, %)
Papilledema (N, %)
Measurable M protein (N, %)
IgA-type M protein (N, %)
Edema (N, %)
Ascites (N, %)
Respiratory symptoms (N, %)
Pulmonary hypertension (N, %)
Osteosclerosis (N, %)
Serum VEGF > 2000 pg/mL (N, %)
28 (41.8)
40 (59.7)
50 (74.6)
49 (73.1)
13 (n = 26) (50.0)
54 (80.6)
6 (9.0)
63 (94.0)
37 (55.2)
46 (68.7)
22 (32.8)
51 (76.1)
67 (100.0)
59 (88.1)
6 (9.0)
19 (28.4)
27 (40.3)
38 (n = 51) (74.5)
113 (48.7)
106 (45.7)
153 (65.9)
141 (60.8)
30 (n = 47) (63.8)
151 (65.1)
36 (15.5)
200 (86.2)
117 (50.4)
124 (53.4)
58 (25.0)
143 (61.6)
191 (82.3)
91 (39.2)
11 (4.7)
35 (15.1)
129 (55.6)
144 (n = 184) (78.3)
0.278
0.045
0.182
0.066
0.250
0.018
0.179
0.093
0.490
0.123
0.203
0.017
0.997
0.000
0.190
0.013
0.019
0.571
27 (41.5)
39 (60.0)
48 (73.8)
48 (73.8)
12 (n = 25) (48.0)
53 (81.5)
6 (9.2)
61 (93.8)
35 (53.8)
44 (67.7)
22 (33.8)
51 (78.5)
65 (100.0)
57 (87.7)
5 (7.7)
18 (27.7)
25 (38.5)
36 (n = 49) (73.5)
24 (36.9)
28 (43.1)
37 (56.9)
36 (55.4)
7 (n = 11) (63.6)
46 (70.8)
14 (21.5)
59 (90.8)
38 (58.5)
33 (50.8)
16 (24.6)
44 (67.7)
56 (86.2)
24 (36.9)
3 (4.7)
9 (13.8)
38 (58.5)
44 (n = 52) (84.6)
0.544
0.055
0.044
0.029
0.387
0.153
0.058
0.513
0.596
0.106
0.249
0.169
0.999
0.000
0.470
0.056
0.007
0.172
Renal manifestations
Microhematuria (N, %)
Proteinuria > 1.0 g (N, %)
13 (19.4)
7 (10.4)
0.003
0.056
13 (20.0)
7 (10.8)
16 (6.9)
10 (4.3)
6 (9.2)
2 (3.1)
0.076
0.084
RI, renal impairment; ONLS, overall neuropathy limitation scale; VEGF, vascular endothelial growth factor.
*If not specified, P-values were calculated using univariate logistic regression analysis.
a
The number of patients with lymph-node biopsy was labeled. Due to the limited number of patients in this variable comparison, χ2 test was used.
278
W. Ye et al.
microhematuria (OR 3.603, 95% CI 1.248–10.401, P = 0.018) were
independently associated with renal impairment.
Because age and sex inherently influence renal function, we
matched these two variables with a propensity score-based
method. This allowed us to compare the two groups (renal impairment versus non-renal impairment) with 65 matched pairs
(overall balance test, P = 0.923) (Table 2). The univariate and
multivariate analyses consistently identified an association between renal impairment and ascites (OR 8.029, 95% CI 2.949–
22.026, P < 0.001).
Time to renal, VEGF and hematological responses
The median time to a renal response was 4.8 months. The
use of a novel-agent-based regimen was associated with a
more rapid renal response compared with the other therapies
administered (median time to response, 2.1 versus 7.3 months;
P = 0.001) (Figure 3). In terms of other responses, the median
times to VEGF and hematological responses were 6.4 and 21.4
months, respectively.
Survival characteristics
After a median follow-up of 27.4 months (range, 0.5–175.5),
33 of 299 patients died (Supplementary data, Table S1), 10
within 3 months of treatment initiation (early death). In the
renal impairment group, 11 patients died (early death, n = 5),
and the causes of death included disease-progression-related
renal failure (n = 6), cardiopulmonary failure (n = 1), tuberculosis (n = 2) and unknown reasons (n = 2). Renal impairment at
baseline was associated with a shortened overall survival (OS)
(5-year OS rate 70.7%), compared with all other patients
(5-year OS rate 87.4%, P = 0.021) and age- and sex-matched patients (5-year OS rate 89.5%, P = 0.042; the median OS was not
reached by all patients) (Figure 4A). When we subdivided the
patients according to their renal function, inferior survival
was observed only in patients with severe renal impairment
(eGFR < 30 mL/min/1.73 m2; 5-year OS rate 69.3%, P = 0.004),
but not in those with moderate dysfunction (eGFR 30–59 mL/
min/1.73 m2; 5-year OS rate 77.5%, P = 0.277), compared with
patients without renal impairment (Figure 4B). Early death
mainly occurred in patients with severe renal impairment
Table 3. Renal, hematological and VEGF responses according to primary therapeutic regimens
2
Baseline eGFR (mL/min/1.73 m ) (median, range)
Best eGFR after treatment (mL/min/1.73 m2) (median, range)
Renal response (N, %)
Median time to renal response (months)
Hematological complete remission (N, %)
VEGF complete remission (N, %)a
Melphalan (n = 14)
ASCT (n = 7)
Novel agent (n = 22)
Others (n = 4)
39 (16–58)
56 (27–110)
7 (50.0)
11.7
3 (21.4)
5 (n = 13) (38.5)
51 (29–56)
78 (51–107)
6 (85.7)
4.8
6 (85.7)
6 (n = 6) (100.0)
39 (15–59)
83 (14–112)
18 (81.8)
2.1
8 (36.4)
11 (n = 19) (57.9)
32 (18–53)
39 (12–58)
0 (0.0)
Not reached
0 (0.0)
0 (n = 2) (0.0)
eGFR, estimated glomerular filtration rate; ASCT, autologous stem cell transplantation; VEGF, vascular endothelial growth factor.
a
The number of patients with measurable serum VEGF before and after treatment was labeled.
Renal impairment in POEMS syndrome
279
ORIGINAL ARTICLE
Treatment and renal response
Patients with renal impairment received the following therapies: melphalan-based regimen (n = 16), autologous stem cell
transplantation (n = 8), novel-agent-based regimen (n = 26) or
others (n = 17). No patient required dialysis.
As renal impairment, rather than massive proteinuria, was
the major renal finding, renal response was determined based
on eGFR changes. Forty-seven patients had at least one additional renal function evaluation at our hospital after treatment,
which was included in the renal response evaluation. A renal
response was observed in 31 patients (66.0%). In terms of the
primary therapy, a renal response was achieved in 50.0% of
the melphalan group, 85.7% of the autologous stem cell transplantation group, 81.8% of the novel agent group and 0.0% of
the others (Table 3).
Patients without a renal response were older (age >55 years,
68.8 versus 35.5%, P = 0.037) and had a longer time from symptom onset to treatment (time from onset to treatment >1 year,
93.8 versus 58.1%, P = 0.030). In contrast, a renal response was
more commonly observed in patients receiving a novel-agentbased therapy (58.1 versus 25.0%, P = 0.037), who achieved
hematological remission (51.6 versus 6.3%, P = 0.011) and
VEGF remission (66.7 versus 20.0%, P = 0.018) (Table 4). After
adjustment in the multivariate analysis, the time interval between
symptom onset and treatment >1 year (OR 0.059, 95% CI 0.004–
0.910, P = 0.043) and VEGF remission (OR 15.958, 95% CI
1.005–253.516, P = 0.050) were statistically significantly associated with a renal response. Because the borderline significance
of the association between VEGF remission and a renal response
might be related to the unavailability of post-treatment VEGF
data for seven patients, we used a receiver-operating characteristic
analysis to determine whether certain reductions in VEGF
translated into a renal response. The best VEGF reduction ensuring a renal response was 85.8% (area under curve 0.85, 95% CI
0.720–0.980, P = 0.007), with a sensitivity of 70% and a specificity
of 100%.
Among the whole retrospective cohort, one patient had renal
impairment at baseline and also later disease relapse. Associated
changes in his/her renal function and VEGF levels were observed
during the course of disease (Figure 2). A renal response (eGFR
increased from 16 to 52 mL/min/1.73 m2) and a reduction in
VEGF (serum levels decreased from 3894 to 615 pg/mL) were
achieved after the first treatment with a melphalan-based regimen. However, the patient’s disease relapsed, with the recurrence
of peripheral edema and ascites (eGFR 36 mL/min/1.73 m2 and
serum VEGF 4439 pg/mL), and a lenalidomide-based regimen
was given. On the patient’s most recent visit, eGFR was 55 mL/
min/1.73 m2 and serum VEGF was 3249 pg/mL and was still
receiving treatment.
ORIGINAL ARTICLE
Table 4. Characteristics of patients with and without renal response
Demographic information
Age > 55 years (N, %)
Male (N, %)
Symptom to treatment >1 year (N, %)
POEMS features
ONLS ≥ 4 (N, %)
Lymphadenopathy (N, %)
Hepatomegaly (N, %)
Splenomegaly (N, %)
Castleman’s disease (N, %)a
Hypothyroidism (N, %)
Diabetes (N, %)
Measurable M protein (N, %)
IgA-type M protein (N, %)
Hyperpigmentation (N, %)
Angioma (N, %)
Papilledema (N, %)
Edema (N, %)
Ascites (N, %)
24-h urine protein >1.0 g (N, %)
Microhematuria (N, %)
eGFR < 30 mL/min/1.73 m2 (N, %)
Osteosclerosis (N, %)
Baseline VEGF >2000 pg/mL (N, %)
Treatment and response
Novel-agent-based regimen (N, %)
VEGF complete remission (N, %)
Hematological complete remission (N, %)
No response (n = 16)
Renal response (n = 31)
OR (95% CI)
P-value*
11 (68.8)
11 (68.8)
15 (93.8)
11 (35.5)
17 (54.8)
18 (58.1)
0.250 (0.069–0.906)
0.552 (0.155–1.969)
0.092 (0.011–0.789)
0.037
0.360
0.030
8 (50.0)
12 (75.0)
8 (50.0)
13 (81.3)
3 (50.0) (n = 6)
11 (68.8)
2 (12.5)
6 (37.5)
10 (62.5)
16 (100.0)
8 (50.0)
10 (62.5)
16 (100.0)
16 (100.0)
2 (12.5)
2 (12.5)
6 (37.5)
10 (62.5)
8 (n = 13) (61.5)
17 (54.8)
24 (77.4)
19 (61.3)
23 (74.2)
6 (54.5) (n = 11)
29 (93.5)
4 (12.9)
13 (41.9)
25 (80.6)
30 (96.8)
21 (67.8)
24 (77.4)
31 (100)
25 (80.6)
2 (6.5)
9 (29.0)
4 (12.9)
12 (38.7)
25 (n = 30) (83.3)
1.214 (0.363–4.066)
1.143 (0.279–4.683)
1.583 (0.469–5.350)
0.663 (0.149–2.947)
NA
6.591 (1.111–39.108)
1.037 (0.169–6.375)
1.204 (0.349–4.152)
2.500 (0.649–9.629)
NA
2.100 (0.610–7.227)
2.057 (0.551–7.674)
NA
NA
0.743 (0.260–4.183)
2.864 (0.538–15.247)
0.690 (0.145–1.805)
0.379 (0.109–1.314)
3.125 (0.716–13.635)
0.753
0.853
0.459
0.590
1.000
0.038
0.969
0.769
0.183
1.000
0.239
0.283
NA
0.999
0.481
0.218
0.060
0.126
0.130
4 (25.0)
2 (n = 10) (20.0)
1 (6.3)
18 (58.1)
20 (n = 30) (66.7)
16 (51.6)
4.154 (1.090–15.827)
8.000 (1.425–44.920)
16.000 (1.876–136.441)
0.037
0.018
0.011
ONLS, overall neuropathy limitation scale; VEGF, vascular endothelial growth factor; eGFR, estimated glomerular filtration rate; NA, not applicable.
*If not specified, P-values were calculated using univariate logistic regression analysis.
a
The number of patients with lymph-node biopsy was labeled. Due to the limited number of patients in this variable comparison, Fisher’s exact test was used.
F I G U R E 2 : Changes in eGFR and serum VEGF levels during the
course of the disease in a patient with baseline renal impairment and
later disease relapse. MDex, melphalan plus dexamethasone regimen;
LDex, lenalidomide plus dexamethasone regimen.
(26.6%) rather than in those with moderate (1.9%, P = 0.001) or
no renal impairment (2.2%, P < 0.001).
We then analyzed the survival characteristics of patients
with a renal response and found that the survival of patients
who achieved a renal response (5-year OS rate 100.0%) was
similar to that of all other patients (5-year OS rate 87.4%, P =
0.099) and to that of age- and sex-matched patients (5-year OS
rate 89.5%, P = 0.141). However, inferior survival was observed when there was no renal response (5-year OS rate 62.5%,
P < 0.001) (Figure 4C). In a 3-month landmark analysis,
280
F I G U R E 3 : Times to renal responses in patients treated with and
without a novel-agent-based regimen.
inferior survival was still observed in patients without renal response (5-year OS rate 83.3%), compared with those who
achieved renal response (P = 0.025), but not in those without
renal impairment at baseline (5-year OS rate 89.8%, P = 0.231)
(Figure 4D).
W. Ye et al.
DISCUSSION
Our series is the first comprehensive study of both renal manifestations and related laboratory findings in a large cohort of patients with POEMS syndrome. Besides the frequently reported
hematuria and proteinuria, we found that impaired renal function was common, with a prevalence of 22.4% in this cohort.
Of these patients with impaired renal function, 77.6% had moderately reduced eGFR (30–59 mL/min/1.73 m2), and renal function was restored after treatment in 66.0%.
Manifestations of renal impairment in patients with POEMS
syndrome are protean, and there is no consensus of its definition in these patients. Massive proteinuria is uncommon,
both in our current cohort (5.7%) and in the well-described
Mayo series (5%), making criteria developed for light chain
amyloidosis inapplicable [15, 16]. In contrast, the decreased
glomerular filtration rate was relatively common. We calculated
eGFR with the use of stabilized creatinine levels at diagnosis,
hoping to reflect the renal function more accurately, as the absolute level of serum creatinine may depend on factors including age, sex and muscle mass. Borrowing the definition of renal
impairment from multiple myeloma, we found that 22.4%
of POEMS patients in the current cohort had baseline renal impairment, which is much higher than the prevalence in Mayo series (2%, defined by serum creatinine level >1.5 mg/dL) [15]. Two
reasons may explain this discrepancy. First, POEMS syndrome is
commonly misdiagnosed in China due to its rarity and complexity, and our patients often had a long time interval between
symptom onset and correct diagnosis [17]. Renal impairment
Renal impairment in POEMS syndrome
could result from long-term course of disease without intervention. Secondly, renal impairment is strongly associated with ascites in our analyses, and there is a much higher frequency of
ascites in POEMS patients from China (50–55%) and Japan
(52%), compared with the Mayo series (7%) [17, 18].
The exact mechanism of renal impairment in POEMS syndrome is unclear. Therefore, we investigated this issue based
on several analyses. Ascites was shown to be a factor strongly
associated with renal impairment in both the univariate and
multivariate models, before and after the matching procedure.
This common complication of POEMS syndrome has features consistent with exudation and is thought to result from
cytokine-induced vascular hyperpermeability [19]. However,
baseline levels of serum VEGF, the key pathogenic cytokine in
POEMS syndrome, which can cause microvascular leakage, did
not differ between patients with and without renal impairment.
This was not surprising, because the absolute level of VEGF is
not associated with the severity of POEMS syndrome. In contrast,
although clinical variables correlate with disease severity to different extents, they can predict certain clinical consequences, as
shown in studies of peripheral blood stem cell mobilization in
patients with POEMS syndrome [20, 21]. Ascites could be the
result of VEGF-induced microvascular hyperpermeability and
is regarded as a clinical surrogate for disease severity, which is
related to renal impairment. This consideration was confirmed
in our renal response analysis. A shorter disease course before
treatment and VEGF remission were associated with a renal
response. The translation of a reduction in VEGF to a renal
response was further corroborated by a receiver-operating characteristic analysis and the parallel time courses of the renal and
281
ORIGINAL ARTICLE
F I G U R E 4 : Survival characteristics. OS in patients with baseline renal impairment (A) and patients subdivided into moderate and severe renal
dysfunction groups (B). (C) OS of treated patients with and without a renal response. (D) Three-month landmark analysis.
ORIGINAL ARTICLE
VEGF responses. Although renal response seemed to be faster
than VEGF response (median time, 4.8 versus 6.4 months), it
was mainly attributed to the different measurement intervals, rather than a real preceding renal response. Serum creatinine,
which can be easily measured in clinical laboratory, was monitored more frequently than serum VEGF, which was measured
every 1–3 months in the early course of treatment.
The renal pathological findings also revealed glomerular
changes, with marked endothelial proliferation and capillary collapse, consistent with previous studies [3–6], which could result
from VEGF stimulation. The marked recovery of these lesions,
with a concomitant reduction in serum VEGF, has been reported
previously in patients after effective therapy [5]. All these findings support the role of VEGF in the renal impairment associated
with POEMS syndrome and also emphasize the concept that the
trends in serum VEGF, rather than its absolute levels, correlate
with disease changes and clinical benefit. The detailed mechanism of how VEGF triggers glomerular changes, and the consequent renal dysfunction, requires further study.
Another finding with significance for clinical practice is the
effectiveness of novel-agent-based regimens in reversing renal
impairment in POEMS syndrome. Of the three novel agents examined, lenalidomide was most frequently used in our patients
(n = 95, accounting for 31.8% of the entire cohort and 94.1% of
the novel-agent group). Although we and others have had successful experiences in treating POEMS patients with bortezomib (n = 3) and thalidomide (n = 3), including patients with
severe renal impairment [22–24], we are still concerned about
their potential adverse effects, including drug-induced peripheral neuropathy, which may complicate the management of
patients with POEMS syndrome. Therefore, we do not regularly
use these agents in our daily practice. The results for the novel
agents in this study mainly pertain to lenalidomide. This
second-generation immunomodulatory agent has been administered to POEMS patients as a first- or second-line treatment
with promising efficacy in both symptomatic improvement and
VEGF reduction [25]. However, its use in POEMS patients with
renal impairment has not been investigated. Our data show, for
the first time, that lenalidomide not only produced a notable recovery of renal function (81.8%), but also achieved the response
more rapidly, which parallels the results achieved in myeloma
patients with renal impairment [26, 27]. It should be mentioned that the lenalidomide regimen requires the dose to be
adjusted, according to the patient’s renal function [28]. In addition, the successful restoration of renal function reversed the
adverse prognosis associated with renal impairment, similar
to the findings in myeloma studies [29, 30]. Sustained renal impairment is associated with an increased risk of early death due
to various complications, observed both in our patients without
renal response and in myeloma kidney patients [13]. Therefore,
we performed a 3-month landmark study to evaluate outcomes
in patients who lived long enough to benefit from anti-plasma
cell therapy, which has become an essential part of survival analysis in patients with severe end-organ dysfunction at baseline
[31, 32]. An improved survival was still observed in patients
who achieved a renal response versus those who did not. Meanwhile, a similar long-term survival was noted in patients without renal response versus those without renal impairment at
282
baseline. These findings should not be over-interpreted, because of the event lost and small number of patients remaining
in the landmark cohort, which limited the statistical power.
Moreover, inferior long-term survival was also observed in
patients with severe renal impairment at baseline, which was
also associated with a more common early death. These patients
tended to have poor renal response (P = 0.060), even in this era
of novel agents. Therefore, more effective interventions should
be explored, in the hope of reversing advanced renal impairment and controling early death.
The retrospective nature of our study was an intrinsic limitation. First, the renal response was not evaluated in all patients
with renal impairment at baseline. Because our patients were
referred from more than 30 centers throughout the country,
it was difficult to regularly follow-up all the patients at our single institution after the treatment regimen was prescribed. This
shortcoming could be resolved if a national network of POEMS
syndrome were established in China. Secondly, the number of
patients with renal biopsy was highly limited, and indications
for this procedure were also not well defined, mainly based
on the treating physicians’ judgments. However, pathological
findings in the current study were consistent with previous reports and showed unique features for POEMS syndrome [3–6].
These alterations, based on previous studies and our current
analyses, could be principally attributed to VEGF stimulation.
Thirdly, the association between reduced VEGF and a renal response does not imply a strict causal relationship. This must be
carefully assessed in prospective studies, with response evaluation at carefully designed time points to delineate their response dynamics.
In summary, renal impairment is a common complication
in patients with POEMS syndrome. It can be reversed in most
patients receiving an effective therapy, especially a novel-agentbased regimen. The restoration of renal function should improve the long-term clinical outcomes of patients, especially
those with only moderate renal impairment. These observations require further validations.
S U P P L E M E N TA R Y D ATA
Supplementary data are available online at http://ndt.oxford
journals.org.
CONTRIBUTIONS
W.Y. participated in study design and reviewed the renal pathology. C.W. was involved in study design, collected the data,
analyzed the data and wrote the manuscript. H.C. measured
the serum VEGF. Q.-Q.C., M.-H.D. and X.-X.C. participated
in patient recruitment. H.L. reviewed the renal pathology and
revised the manuscript. D.-B.Z. supervised the study and analyzed the data. J.L. designed the study, recruited the patients, analyzed the data, wrote the manuscript and critically revised the
manuscript. All authors have approved the final manuscript.
W. Ye et al.
AC K N O W L E D G E M E N T S
The authors would like to thank all the patients who participated in this study. We would also like to extend our appreciation to Tianjiao Li for her maintenance of the serum sample
collection from patients with POEMS syndrome. The Beijing
Natural Science Foundation (no. 7142130), the Specialized
Research Fund for the Doctoral Program of Higher Education
(no. 2013110611000), the Capital Health Research and
Development of Special (no. 2011-4001-03) and Peking
Union Medical College New Star (2011, for LJ) supported
this research.
C O N F L I C T O F I N T E R E S T S TAT E M E N T
The authors declare that they have no conflicts of interest.
REFERENCES
Renal impairment in POEMS syndrome
Received for publication: 16.3.2015; Accepted in revised form: 3.6.2015
283
ORIGINAL ARTICLE
1. Li J, Zhou DB. New advances in the diagnosis and treatment of POEMS
syndrome. Br J Haematol 2013; 161: 303–315
2. Dispenzieri A. POEMS syndrome: 2014 update on diagnosis, riskstratification, and management. Am J Hematol 2014; 89: 214–223
3. Navis GJ, Dullaart RP, Vellenga E et al. Renal disease in POEMS syndrome:
report on a case and review of the literature. Nephrol Dial Transplant 1994;
9: 1477–1481
4. Nakamoto Y, Imai H, Yasuda T et al. A spectrum of clinicopathological
features of nephropathy associated with POEMS syndrome. Nephrol Dial
Transplant 1999; 14: 2370–2378
5. Sanada S, Ookawara S, Karube H et al. Marked recovery of severe renal
lesions in POEMS syndrome with high-dose melphalan therapy supported
by autologous blood stem cell transplantation. Am J Kidney Dis 2006; 47:
672–679
6. Nakamura Y, Nishimura M, Terano T et al. A patient with POEMS syndrome: the pathology of glomerular microangiopathy. Tohoku J Exp Med
2013; 231: 229–234
7. Li J, Zhang W, Jiao L et al. Combination of melphalan and dexamethasone
for patients with newly diagnosed POEMS syndrome. Blood 2011; 117:
6445–6449
8. Levey AS, Stevens LA, Schmid CH et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009; 150: 604–612
9. Wang C, Zhou YL, Cai H et al. Markedly elevated serum total N-terminal
propeptide of type I collagen is a novel marker for the diagnosis and followup of patients with POEMS syndrome. Haematologica 2014; 99: e78–e80
10. National Kidney Foundation. K/DOQI clinical practice guidelines for
chronic kidney disease: evaluation, classification, and stratification. Am J
Kidney Dis 2002; 39(Suppl. 1): S1–S266
11. Dimopoulos MA, Terpos E, Chanan-Khan A et al. Renal impairment in patients with multiple myeloma: a consensus statement on behalf of the International Myeloma Working Group. J Clin Oncol 2010; 28: 4976–4984
12. Ludwig H, Adam Z, Hajek R et al. Light chain-induced acute renal failure
can be reversed by bortezomib–doxorubicin–dexamethasone in multiple
myeloma: results of a phase II study. J Clin Oncol 2010; 28: 4635–4641
13. Dimopoulos MA, Roussou M, Gkotzamanidou M et al. The role of
novel agents on the reversibility of renal impairment in newly diagnosed
symptomatic patients with multiple myeloma. Leukemia 2013; 27: 423–429
14. D’Souza A, Lacy M, Gertz M et al. Long-term outcomes after autologous
stem cell transplantation for patients with POEMS syndrome (osteosclerotic myeloma): a single-center experience. Blood 2012; 120: 56–62
15. Dispenzieri A, Kyle RA, Lacy MQ et al. POEMS syndrome: definitions and
long-term outcome. Blood 2003; 101: 2496–2506
16. Gertz MA, Comenzo R, Falk RH et al. Definition of organ involvement and
treatment response in immunoglobulin light chain amyloidosis (AL): a
consensus opinion from the 10th International Symposium on Amyloid
and Amyloidosis, Tours, France, 18–22 April 2004. Am J Hematol 2005;
79: 319–328
17. Li J, Zhou DB, Huang Z et al. Clinical characteristics and long-term outcome of patients with POEMS syndrome in China. Ann Hematol 2011;
90: 819–826
18. Nakanishi T, Sobue I, Toyokura Y et al. The Crow–Fukase syndrome: a
study of 102 cases in Japan. Neurology 1984; 34: 712–720
19. Cui RT, Yu SY, Huang XS et al. The characteristics of ascites in patients with
POEMS syndrome. Ann Hematol 2013; 92: 1661–1664
20. Shimizu N, Nakaseko C, Sakaida E et al. Factors associated with the efficiency of PBSC collection in POEMS syndrome patients undergoing
autologous PBSC transplantation. Bone Marrow Transplant 2012; 47:
1010–1012
21. Li J, Zhang W, Duan MH et al. PBSC mobilization in newly diagnosed
patients with POEMS syndrome: outcomes and prognostic factors. Bone
Marrow Transplant 2013; 48: 233–237
22. Kuwabara S, Misawa S, Kanai K et al. Thalidomide reduces serum VEGF
levels and improves peripheral neuropathy in POEMS syndrome. J Neurol
Neurosurg Psychiatry 2008; 79: 1255–1257
23. Kaygusuz I, Tezcan H, Cetiner M et al. Bortezomib: a new therapeutic
option for POEMS syndrome. Eur J Haematol 2010; 84: 175–177
24. Li J, Zhang W, Kang WY et al. Bortezomib and dexamethasone as first-line
therapy for a patient with newly diagnosed polyneuropathy, organomegaly,
endocrinopathy, M protein and skin changes syndrome complicated by
renal failure. Leuk Lymphoma 2012; 53: 2527–2529
25. Zagouri F, Kastritis E, Gavriatopoulou M et al. Lenalidomide in patients
with POEMS syndrome: a systematic review and pooled analysis. Leuk
Lymphoma 2014; 55: 2018–2023
26. Dimopoulos M, Alegre A, Stadtmauer EA et al. The efficacy and safety
of lenalidomide plus dexamethasone in relapsed and/or refractory multiple
myeloma patients with impaired renal function. Cancer 2010; 116:
3807–3814
27. Klein U, Neben K, Hielscher T et al. Lenalidomide in combination with
dexamethasone: effective regimen in patients with relapsed or refractory
multiple myeloma complicated by renal impairment. Ann Hematol 2011;
90: 429–439
28. Chen N, Lau H, Kong L et al. Pharmacokinetics of lenalidomide in subjects
with various degrees of renal impairment and in subjects on hemodialysis.
J Clin Pharmacol 2007; 47: 1466–1475
29. Martina K, Gabriele I, Josefina U et al. Use of lenalidomide may surmount
the adverse prognosis induced by renal impairment in patients ( pts) with
relapsed/refractory multiple myeloma (MM). Blood (ASH Annual Meeting
Abstracts) 2010; 116: Abstract 3042
30. Eleftherakis-Papapiakovou E, Kastritis E, Roussou M et al. Renal impairment is not an independent adverse prognostic factor in patients with multiple myeloma treated upfront with novel agent-based regimens. Leuk
Lymphoma 2011; 52: 2299–2303
31. Kastritis E, Wechalekar AD, Dimopoulos MA et al. Bortezomib with or
without dexamethasone in primary systemic (light chain) amyloidosis.
J Clin Oncol 2010; 28: 1031–1037
32. Jaccard A, Comenzo RL, Hari P et al. Efficacy of bortezomib, cyclophosphamide and dexamethasone in treatment-naïve patients with high-risk cardiac AL amyloidosis (Mayo Clinic stage III). Haematologica 2014; 99:
1479–1485