retrospective analysis of 37 287 observation years after peripheral blood stem cell donation

Highlights  Retrospective survey of 15,445 individuals who donated peripheral blood stem cells PBSC or bone marrow BM between 1992 and 2009  Almost 95% of responders assessed their hea

Accepted Manuscript

Title: Retrospective Analysis of 37,287 Observation Years after Peripheral

Blood Stem Cell Donation

Author: Alexander H Schmidt, Thilo Mengling, Camila J

Hernández-Frederick, Gabi Rall, Julia Pingel, Johannes Schetelig, Gerhard Ehninger

Please cite this article as: Alexander H Schmidt, Thilo Mengling, Camila J

Hernández-Frederick, Gabi Rall, Julia Pingel, Johannes Schetelig, Gerhard Ehninger, Retrospective Analysis

of 37,287 Observation Years after Peripheral Blood Stem Cell Donation, Biology of Blood and

Marrow Transplantation (2017), http://dx.doi.org/doi: 10.1016/j.bbmt.2017.02.014

This is a PDF file of an unedited manuscript that has been accepted for publication As a service

to our customers we are providing this early version of the manuscript The manuscript will

undergo copyediting, typesetting, and review of the resulting proof before it is published in its

final form Please note that during the production process errors may be discovered which could

affect the content, and all legal disclaimers that apply to the journal pertain

Retrospective analysis of 37,287 observation years after peripheral blood stem cell

donation

Alexander H Schmidt, MD, PhD1*, Thilo Mengling, MD1*, Camila J Hernández-Frederick

PhD1, Gabi Rall1, Julia Pingel, PhD1, Johannes Schetelig, MD, MSc2,3, and Gerhard

Ehninger, MD3

2DKMS, Clinical Trials Unit, Dresden, Germany

3Internal Medicine I, University Hospital Carl Gustav Carus, Dresden, Germany

* AHS and TM contributed equally to this work

Short title: Analysis of 37,287 observation years after PBSC

Phone: +49-221-940582-3421

Fax: +49-221-940582-3499

Authors’ contributions

AHS, TM, GR, and GE designed the study TM collected data AHS, TM, CJHF, JP, and GE

analyzed data All authors contributed to data interpretation AHS prepared the manuscript

with support by TM, CJHF, JP, JS and GE All authors revised and approved the manuscript

Financial Disclosure Statement

The authors report no conflict of interest

Keywords: recombinant human granulocyte-colony stimulating factor (rhG-CSF),

hematopoietic stem cell donor, follow-up, peripheral blood stem cells, bone marrow

Word count: Abstract: 217 words

Main text: 3,522 words Number of tables: 6 Number of figures: 1 Supplemental file: 1

Highlights

 Retrospective survey of 15,445 individuals who donated peripheral blood stem cells

(PBSC) or bone marrow (BM) between 1992 and 2009

 Almost 95% of responders assessed their health conditions as very good or good

 No differences in the frequency of reported health events between PBSC and BM

donors

 No evidence that either PBSC or BM donation are associated with increased risks of

malignancies

Abstract

Donor safety is of utmost importance in the setting of hematopoietic stem cell donation

Follow-up is indicated to detect potential long-term risks for donors We sent a follow-up

questionnaire to 15,445 donors of peripheral blood stem cells (PBSC) or bone marrow (BM)

within a retrospective study design The return rate was 91.3% resulting in 37,287

observation years for PBSC donors and 25,656 for BM donors Most donors assessed their

health conditions as very good or good, and had not been hospitalized or received long-term

medical treatment including prescribed medication for more than 4 weeks since donation

While there were no differences in the frequency of reported health events, BM donors more

often rated their general health as very good or good Ninety-five percent of donors after BM

or PBSC donation respectively would consider a second stem cell donation In total, 93

malignancies were reported The standardized incidence ratio (SIR) for a diagnosis of any

type of cancer after PBSC donation was 0.94 (95%-CI, 0.70 - 1.24) with a SIR below 1

indicating a lower risk than in the ageand sex-matched population The SIR for a diagnosis

of leukemia was 0 (95%-CI, 0 - 1.88) In summary, we found no evidence that either PBSC

or BM donation are associated with increased risks of malignancies or other severe health

problems

Introduction

Allogeneic stem cell donors undergo either bone marrow (BM) harvest in general anesthesia

or leukapheresis after mobilization of hematopoietic stem cells with recombinant human

granulocyte-colony stimulating factor (rhG-CSF) Although both procedures are regarded as

safe,1-6 it is common understanding that there is a need for long-term follow-up of large

donor cohorts in order to identify and further minimize potential risks for donors.2, 7-9

Long-term donor follow-up is of special relevance for peripheral blood stem cell (PBSC)

donors as concern was raised, based on experimental results10 or clinical data,11 regarding

potential correlations between short-term rhG-CSF application and the development of

hematological malignancies Increased incidences of very rare events are difficult to prove

for methodological reasons.12

From March 1992 to January 2009, 16,270 stem cell donations of 15,531 donors from DKMS

Germany had been carried out, thereof 11,540 PBSC and 4,730 BM donations

Single-center results of DKMS’ prospective PBSC donor follow-up have been published

before.4 A small but statistically significant lower absolute neutrophil count within the normal

range was observed after the follow-up period of five years in that study Four hematological

malignancies among 12 total cancer diagnoses had been observed: one acute myeloid

leukemia (AML) case, one chronic lymphatic leukemia (CLL) case, and two cases of

Hodgkin lymphoma Statistically, the incidence of Hodgkin lymphoma differed significantly

from the age- and gender-adjusted German population

In this work, we present analyses based on a retrospective follow-up project that included

the mailing of questionnaires to all DKMS donors who had donated PBSC or BM from March

1992 to January 2009 and telephone-based interviews of initial non-responders In our

analyses we especially focused on malignancies, autoimmune disorders, and mental and

psychosocial disorders Malignancies were considered due to the discussion regarding

potential long-term risks of rhG-CSF application.10, 11 Regarding autoimmune disorders, there

is evidence that they may be induced or boosted by rhG-CSF application.13, 14 Positive

psychosocial effects of stem cell donation have been described.15 There is, however, also

potential for negative emotional stress, for example, in the case of patient death after

hematopoietic stem cell transplantation.16

Materials and methods

Donations

An overview of all donations by donors from DKMS Germany between March 1992 and

January 2009 is given in Table 1 Generally, DKMS’ respective policy sets a limit of two

PBSC plus two BM donations per donor It follows from Table 1 that nearly all donors

donated once (95.3%) or twice (4.6%)

The standard mobilization protocol for PBSC donation consisted of daily doses of 7.5 µg/kg

lenograstim for 5-6 days In few cases, daily doses of 10 µg/kg filgrastim or single doses of

12 mg PEG-filgrastim17 were applied BM harvest was carried out under general anesthesia

Follow-up data

Follow-up questionnaires were sent out from December 2008 to February 2009 to all DKMS

donors who had donated PBSC or BM between March 1992 and January 2009 Only

exceptions were known cases of death (n=20) and donors who previously had asked not to

be contacted again or were not contactable for other reasons as, for example, emigration

(n=66) In total, 15,445 donors were contacted (Figure 1) The study was approved by the

Ethics Committee of the Technical University of Dresden, Germany

Donors were asked about general health condition (Question #1, four categories ranging

from “very good” to “reduced”), hospitalization or long-term medical treatment since donation

(Question #2, “yes” or “no”), use of prescription drugs regularly or for more than 4 weeks

since donation (Question #3, “yes” or “no”), and willingness to donate again (Question #4,

four categories from “yes” to “no”) Donors with hospitalization or long-term treatment were

asked to give comments and to make an assignment to one of 11 categories including, for

example, cardiovascular system and malignancies Users of prescribed drugs were asked to

list the drugs The questionnaire is included in the Supplementary Information

A reminder was sent to all donors who did not answer within ≈50 days We tried to contact

initial non-responders (n=2,319) by phone between August 2011 and December 2011 in

order to complete the questionnaire

PBSC donors who donated between January 1996 and January 2008 at the apheresis

center in Dresden are also included in the study by Hölig et al 4

As with any self-report survey, certain limitations to validity are inherent To minimize a

potential bias, we focused our analyses either on conditions that are unlikely to be

underreported and clarified any ambiguous report, or on subjective self-assessment

Question #3 about medication was primarily included to cross-check reported diagnoses

Definitions

Health disorders that were reported under Question #2 were encoded according to the 10th

revision of the International Statistical Classification of Diseases and Related Health

Problems (ICD-10) (Supplementary Information) Malignancies, systemic autoimmune

disorders and health conditions leading to permanent exclusion from further stem cell

donations were clarified by DKMS physicians If necessary for correct classification and

donor consented, medical reports were obtained For example, all reported cases of bladder

cancer were evaluated to distinguish between invasive and non-invasive urothelial

carcinomas Statistical analyses were carried out for malignancies, autoimmune disorders,

and mental and psychosocial disorders Definitions of the three disease groups are given in

the Supplementary Information

Statistical analyses

2 tests were used for univariate significance testing Binary logistic regression analyses

were performed with SPSS (version 21.0) (IBM, Armonk, NY, USA) For general answers to

Questions #1-4, p values below 0.01 were regarded as significant due to large sample sizes

and multiple testing For adverse events a ‘test-wise’ significance level of 5% was chosen

Numbers of expected cases for various malignancies in the donor samples (PBSC donors,

BM donors, PBSC+BM donors) were calculated based on age- and gender-adjusted

malignancy incidences of the German population.18, 19 Standard incidence ratios (SIRs) and

corresponding 95% confidence intervals (CI) based on the Poisson distribution were

calculated according to Estève et al.20 This approach is based on the assumption that

potential increases of malignancy risks after PBSC or BM donation are equally distributed

over time

Results

Return rates

In total, 14,094 donors returned the questionnaire in written form or answered questions on

the phone including signed or verbal informed consent Return rates were 91.3% (all

donors), 91.1% (PBSC donors), 91.5% (BM donors), and 92.4% (donors of both PBSC and

BM) The total observation period was 64,933 donor years (37,287 for PBSC donors, 25,656

for BM donors, and 1990 for donors of both PBSC and BM) Characteristics of responding

donors are given in Tables 2 and 3

1,351 (8.7%) donors were non-responders as they did not return the questionnaire, did not

give informed consent or declined to participate in the study

A comparison between responders (n=14,094) and non-responders (n=1,351) showed a

significantly higher number of young (18-40 years) and male donors among the

non-responders (2 tests, p<0.001)

In order to check for a potential bias caused by donors who died in the interval between

donation and the mailing of the questionnaire, we analyzed a sample of 140 non-responding

donors 36 of these donors were in contact with DKMS for various reasons after the

questionnaire had been sent out, 43 could be reached by phone, and for the remaining 61

donors registration office inquiries provided no evidence that they were not alive Taken

together, there was no evidence for a considerable bias due to donor deaths between

donation and the mailing of the questionnaire

General results – univariate analysis

Answers provided by study donors are shown in Table 3 Most donors assessed their health

conditions as very good or good (94.9%), had not been hospitalized or received long-term

medical treatment since donation (82.4%), and neither used prescription drugs regularly nor

had used them for more than four weeks since donation (77.7%) Moreover, the majority of

donors (95.0%) would be willing (“yes” or “probably yes”) to donate again if asked to do so

PBSC and BM donors showed no significant differences regarding their general health

conditions (Question #1, 2 test, p=0.06) For this analysis, pre-defined answers were

combined (very good/good vs moderate/reduced)

Contrary to Question #1 that asked for subjective assessments, Questions #2

(hospitalization or long-term treatment) and #3 (prescription drugs) focused on specific

indications for health-related problems For these questions, answers of PBSC and BM

donors differed significantly (2 tests, p<0.001), with less PBSC donors reporting

health-related problems Answers to Questions #2 and #3 also differed significantly between male

and female donors, between younger (18-40 years) and older (≥41 years) donors, and

between donors with more (<5 years) and less (>5 years) recent donations (2 tests,

p<0.001) Male donors, younger donors and donors with more recent donations less often

reported health-related problems The PBSC sample included significantly more male,

younger (18-40 years) and more recent (<5 years) donors than the BM group (2 tests,

p<0.001) The strong effect of time since donation may result from the fact that Questions #2

and #3 refer to incidents in the interval since donation

Willingness to donate again (Question #4) did not differ significantly between PBSC and BM

donors (2 test, p=0.30) PBSC+BM donors were significantly less often willing to donate

again (2 tests, p<0.001) However, the fact that 89.8% of these donors who already donated

at least twice were willing to donate again is remarkable Similar to the analysis of Question

#1, pre-defined answers were combined for analysis (yes/probably yes vs probably not/no)

General results – multivariate analysis

In multivariate regression analysis, there were no significant differences between PBSC, BM

and PBSC+BM donors with respect to Questions #2-4 (Table 4) Expectedly, odds ratios

(OR) for Questions #2 and #3 essentially show a continuous increase of reported

health-related problems with increasing donor age and increasing observation periods In spite of

similar frequencies of reported health-related problems in PBSC and BM donors, BM donors

assessed their general health condition significantly more often as very good or good than

PBSC donors (OR=1.58, p<0.001) This finding is even more apparent when the subset of

donors is analyzed that did not report a specific health issue or medication In this group,

PBSC donors reported a ‘moderate’ or ‘reduced’ health status with an OR of 2.94 (p<0.001)

compared to BM donors Females assessed their health less often as very good/good

(OR=1.55, p=0.009) Time since donation had no impact on self-assessment

Malignancies

93 malignancies were reported by 90 donors within the study, thereof 56 malignancies of 53

PBSC donors, 33 of BM donors, and four of donors of both PBSC and BM (Supplementary

Table S1 includes a complete list of cases) Six hematological malignancies were reported:

two cases of Hodgkin lymphoma (both in PBSC donors), plasmocytoma (PBSC donor), AML

(BM donor), Non-Hodgkin lymphoma (BM donor), and CLL (donor of both PBSC and BM)

There was no significant difference between PBSC and BM donors in multivariate analysis

(OR=1.08, 95% CI: [0.64, 1.80]) However, we observed a significant difference between

female donors and male donors (OR=1.64, 95% CI: [1.07, 2.52]) Higher donor age also was

a significant risk factor for the occurrence of malignancies

Table 6 shows observed and expected numbers for various malignancies and resulting SIR

values with 95% CI We obtained significantly increased SIR values (lower bound of

95% CI >1) for 2 combinations of diseases or disease groups and donation methods: CLL in

BM+PBSC donors (one case observed, 0.02 expected, SIR=51.44, 95% CI: [1.54, 286.51])

and malignant neoplasms of unspecified female genital organs in PBSC donors (one case

observed, 0.02 expected, SIR=41.13, 95% CI: [1.23, 229.09])

A trend to a lower rate of malignancies compared to the age- and gender-adjusted

malignancy incidences of the German population was found for all donors (SIR 0.84,

95%-CI: [0.68 – 1.03]) The SIR of a cancer diagnosis was lower than expected (SIR 0.7, 95%-95%-CI:

[0.48 – 0.99]) for BM donors but not for PBSC donors (SIR 0.94, 95%-CI: [0.70 – 1.24])

Also, no increased risk was observed for the diagnosis of leukemia (ICD-10: C91-C95) after

PBSC donation (no case observed, 1.60 cases expected, SIR=0.00, 95% CI: [0.00, 1.88])

As indicated by the 95% confidence interval an increase in the risk of leukemia after PBSC

donation by more than 1.88 times can be excluded with a 2.5% error probability based on

this analysis With the same argument, a twofold or higher increase of the risk for

Non-Hodgkin lymphoma (ICD-10: C82-C85; no case observed, 2.17 cases expected, SIR=0.00,

95% CI: [0.00, 1.38]), a seven-fold or stronger increase of the risk for Hodgkin lymphoma

(ICD-10: C81; two cases observed, 1.07 cases expected, SIR=1.87, 95% CI: [0.22, 6.74]),

and a 14-fold or stronger increase of the risk for plasmocytoma (ICD-10: C90; one case

observed, 0.41 cases expected, SIR=2.45, 95% CI: [0.07, 13.65]) can be rejected for PBSC

donors at the same significance level all under the assumption that additional malignancy

cases were evenly distributed over time

SIRs with an upper bound of the 95% CI below 1 were observed for 6 combinations of

malignancies and donation methods: malignant neoplasms of the skin other than melanoma

in all donors (7 cases observed, 19.13 expected, SIR=0.37, 95% CI: [0.15, 0.75]) and in BM

donors (one case observed, 8.51 expected, SIR=0.12, 95% CI: [0.00, 0.65]), lung cancer in

all donors (one case observed, 7.39 expected, SIR=0.14, 95% CI: [0.00, 0.75]) and in BM

donors (no case observed, 3.38 expected, SIR=0.00, 95% CI: [0.00, 0.89]), malignant

neoplasms of lips, oral cavity and pharynx in all donors (no case observed, 5.03 expected,

SIR=0.00, 95% CI: [0.00, 0.60]), and all malignancies excluding malignant neoplasms of the

skin other than melanoma in BM donors (32 cases observed, 45.41 expected, SIR=0.70,

95% CI: [0.48, 0.99])

Autoimmune disorders

106 autoimmune disorders were reported by 102 donors (Supplementary Table S2) PBSC

donors reported significantly less often autoimmune disorders than BM donors (OR=0.60,

95% CI: [0.39, 0.94]) Besides, autoimmune disorders were reported significantly more often

by female compared to male donors (OR=1.78, 95% CI: [1.19, 2.64])

Mental and psychosocial disorders

371 donors reported mental or psychosocial disorders (Supplementary Table S2) or the use

of psychiatric medication There were no significant differences between PBSC and BM

donors (OR=1.03, 95% CI: [0.81, 1.33]) Female donors significantly more often reported

mental or psychosocial disorders or the use of psychiatric medication (OR=2.27, 95% CI:

[1.84, 2.79]) A negative effect of multiple donations could not be observed (OR=1.01, 95%

CI: [0.55, 1.87])

Discussion

In this work, we present results of a retrospective follow-up study based on an

easy-to-complete questionnaire that was mailed to 15,445 individuals who had previously donated

PBSC or BM Most donors reported very good or good general health conditions and would

donate again if asked to do so

In multivariate analysis, no significant differences regarding health-related problems between

PBSC and BM donors could be observed The use of prescription drugs was less often

reported by PBSC donors but significance was not reached Differences between PBSC and

BM donors with respect to health-related problems could be explained by differences in

eligibility criteria for PBSC and BM donation For example, donors with autoimmune

disorders such as autoimmune hypothyroidism, vitiligo or alopecia areata were excluded

from PBSC but not from BM donation, donors with spine affections from BM, but might be

allowed for PBSC

Though they did not report health-related problems significantly less often than PBSC

donors, BM donors assessed their general health conditions significantly more often as very

good or good The even more pronounced correlation of worse health assessment with

PBSC donation in the subset of donors who did not report any specific health issue or

medication indicates that different non-eligibility criteria cannot explain this result, as most of

these pre-existing conditions would have been replied to questions # 2 and 3 In this context,

it may be relevant that the final choice about the donation method is made by the donor We

know from donor center practice that about 10% of the potential donors consent only to one

of the procedures with a clear majority of these donors preferring PBSC donation Therefore,

both donor groups may differ systematically with respect to health-related attitudes Such

differences could potentially cause varying subjective assessments of the general health

status without detectable differences regarding indications for health-related problems

However, the reason for the worse subjective assessment of the health condition by PBSC

donors remains obscure.21 More research is needed in order to delineate the underlying

cause

One donor safety issue lies in the question if short-term application of rhG-CSF may lead to

an increased risk to develop hematological malignancies.5-8, 10 We observed no significant

deviations from expected incidences for hematological malignancies of PBSC or BM donors

in our study This holds also for Hodgkin lymphoma for which a significant increase was

observed before.4 This increase was based on two cases that are also included in our

analysis As the observation period of PBSC donors has increased from 8,234 to 37,287

donor years without additional cases of Hodgkin lymphoma, significance has been lost

Given the upper limit of the 95% confidence interval for the SIR for leukemia after PBSC

donation of 1.88, an increase above factor two is highly unlikely based on this analysis

The two significantly increased SIR values (Table 5) – CLL in PBSC+BM donors and

malignant neoplasms of unspecified female genital organs in PBSC donors – were induced

by one observed case and are, therefore, hardly interpretable

A significant decrease was observed for incidences of lung cancer, malignant neoplasms of

lip, oral cavity and pharynx, and malignant neoplasms of the skin other than melanoma A

correlation between malignancy development and a lack of health-conscious behavior

(smoking, misuse of alcohol) is well-known for lung cancer and malignant neoplasms of lips,

oral cavity and pharynx One might, therefore, hypothesize that stem cell donors show more

often health-conscious behavior than the general population

Most malignancy incidences that were determined in our study lie within ranges that would

be expected from epidemiological data of the German population.18 Theoretically, this result

could be a combined effect of increased malignancy risks through stem cell donation and

underreporting of occurring malignancies However, increased malignancy risks after BM

donation have not been discussed in the literature and do not seem to be plausible The

same holds for correlations of many non-hematological malignant diseases with PBSC

donation Therefore, we conclude that the SIRs obtained in our study are not heavily

affected by underreporting

In our study, PBSC donors reported significantly less often autoimmune disorders than BM

donors It does not seem plausible that the incidence or severity of autoimmune disorders

are increased by BM donation or reduced by PBSC donation Therefore, this result most

probably reflects the stricter eligibility criteria regarding some autoimmune disorders – as, for

example, autoimmune thyroiditis – for PBSC donors