Pediatric Nephrology

https://doi.org/10.1007/s00467-017-3846-6

ORIGINAL ARTICLE

Mycophenolate mofetil following glucocorticoid treatment

in Henoch-Schönlein purpura nephritis: the role of early initiation
and therapeutic drug monitoring
Agnes Hackl 1 & Jan U. Becker 2 & Lisa M. Körner 1 & Rasmus Ehren 1 & Sandra Habbig 1 &
Eva Nüsken 1 & Kai-Dietrich Nüsken 1 & Kathrin Ebner 1 & Max C. Liebau 1,3,4 &
Carsten Müller 5 & Martin Pohl 6 & Lutz T. Weber 1

Received: 29 May 2017 / Revised: 2 November 2017 / Accepted: 3 November 2017

# IPNA 2017

Abstract
Background Henoch-Schönlein purpura (HSP) is the most common vasculitis in childhood and traditionally considered as a self-
limiting disease. However, renal involvement can unfavorably determine long-term prognosis. The reported regimens to treat
HSP nephritis (HSPN) are diverse, indicating that the most effective treatment remains controversial.
Methods This retrospective, single-center study involved 18 patients presenting with HSPN and nephrotic-range proteinuria. We
aimed to investigate the efficacy and safety of mycophenolate mofetil (MMF) and identify a cut-off level for estimated myco-
phenolic acid area under the curve (eMPA-AUC0-12h) values, which can predict complete remission with high sensitivity.
Results Despite prior insufficient therapeutic response to corticosteroids, 89% of patients showed a significant decrease in
proteinuria after 1 month of MMF treatment. None of them relapsed during treatment; however, two children relapsed after
discontinuation. Based on results of a receiver operating characteristic (ROC) analysis, an eMPA-AUC0–12h >56.4 mg*h/l was a
predictor for complete remission within 3 months (80% sensitivity, 83.3% specificity, p = 0.035). During MMF administration,
we encountered no adverse event requiring discontinuation of treatment.
Conclusion Our study demonstrates that MMF is a safe and potentially effective secondary treatment option for children with
HSPN to achieve and maintain long-term remission without serious side effects. To achieve complete remission within 3 months,
resolve severe inflammatory glomerular lesions, and avoid progression to chronic kidney disease, we propose timely diagnosis
and early initiation of MMF with an eMPA-AUC0–12h value of 56.4 mg*h/l.

Electronic supplementary material The online version of this article Keywords Mycophenolic acid . Therapeutic drug monitoring .
(https://doi.org/10.1007/s00467-017-3846-6) contains supplementary Pre-dose level . Cut-off of eMPA-AUC0-12h . Mechanism of
material, which is available to authorized users. action

* Agnes Hackl
agnes.hackl@uk-koeln.de
Abbreviations
1
Pediatric Nephrology, Children’s and Adolescents’ Hospital, ACEi Angiotensin-converting enzyme inhibitors
University Hospital of Cologne, Kerpener Street 62, ARB Angiotensin II receptor blockers
50937 Cologne, Germany BSA Body surface area
2
Institute of Pathology, University Hospital of Cologne, CI Confidence interval,
Cologne, Germany CKD Chronic kidney disease
ESRD End-stage renal disease
3
Nephrology Research Laboratory, Department II of Internal GTP Guanosine-5′-triphosphate
Medicine, University Hospital of Cologne, Cologne, Germany HSP Henoch-Schönlein purpura
4
Center for Molecular Medicine, University Hospital of Cologne, HSPN Henoch-Schönlein purpura nephritis
Cologne, Germany IgACI Immunoglobulin A circulating immune complex
IMPDH Inosine-5′-monophosphate dehydrogenase
5
Department of Therapeutic Drug Monitoring, Centre of IS therapy Immunosuppressive therapy
Pharmacology, University Hospital of Cologne, Cologne, Germany MMF Mycophenolate mofetil
6
Department of General Pediatrics, Adolescent Medicine and MP Methylprednisolone
Neonatology, Medical Center - Faculty of Medicine, University of MPA-C0 Pre-dose level of mycophenolic acid
Freiburg Germany, Freiburg, Germany

eMPA-AUC0-12h Estimated mycophenolic acid–area under the
concentration versus time curve
NA Not applicable
PU Proteinuria
RAAS Renin–angiotensin–aldosterone system
ROC Receiver operating characteristic
TDM Therapeutic drug monitoring
UTI Urinary tract infection
Introduction
Henoch-Schönlein purpura (HSP) is the most common vascu-
litis in childhood, with an incidence of 6.1:100,000 children
<17 years of age and showing a male dominance [1–4]. The
disease is traditionally considered as self-limiting, but renal
involvement, which has been reported in 46% of patients
[5], can unfavorably determine long-term prognosis. In spe-
cialized centers, 20% of patients with renal symptoms prog-
ress to end-stage renal disease (ESRD) [2], constituting 1% of
the pediatric kidney transplant population [6].
The reported regimens of HSP nephritis (HSPN) therapy
are diverse, indicating that the most effective treatment re-
mains controversial. The Kidney Disease: Improving Global
Outcomes (KDIGO) guidelines suggest treatment with
angiotensin-converting enzyme inhibitors (ACEi) or angio-
tensin II receptor blockers (ARB) for patients with proteinuria
>0.5 g/d/1.73m2 body surface area (BSA). Furthermore, in
patients with nephrotic-range proteinuria, orally administered
corticosteroids are recommended for 6 months. Due to lack of
strong evidence, immunosuppressive (IS) therapy is only a
weak suggestion (2D), even in case of nephrotic-range pro-
teinuria with deteriorating kidney function [7]. It is important
to note that this guideline refers to studies conducted in adult
patients with immunoglobulin A (IgA) nephritis.
Additionally, Cochrane analyses, derived from generally
low-quality evidence in adults as well as children, found no
strong evidence of benefit for any of the investigated treatment
approaches in HSPN [8]. However, several retrospective and a
few controlled prospective studies focusing on the pediatric
population with HSPN reported a beneficial role of IS therapy.
Early corticosteroid treatment in prospective studies signifi-
cantly reduced the odds of developing renal disease [9–11].
Furthermore, prompt initiation of prednisolone orally in com-
bination with cyclophosphamide may be beneficial [12].
However, cyclophosphamide alone was not proven to be ef-
fective in a randomized, prospective trial [13]. On the con-
trary, cyclosporin A (CyA) treatment was effective in HSPN
showing nephrotic-range proteinuria, not only in combination
with steroids [14] but also as monotherapy [15]. It was safe
and not inferior to corticosteroid treatment according to a ran-
domized open-label clinical trial [15]. Finally, a regimen of
mizoribine with steroids seemed to be effective in ameliorat-
ing proteinuria and the histological severity of HSPN [16].
Pediatr Nephrol
However, these drugs carry the risk of severe side effects, such
as gonadotoxicity, bone marrow toxicity, nephrotoxicity, neu-
rotoxicity, and hyperuricemia.
Also, for safety reasons in this vulnerable population, my-
cophenolate mofetil (MMF) was proposed as an alternative
therapeutic option. It is a powerful inhibitor of lymphocyte
proliferation [17] and has been successfully used for >20 years
in transplantation medicine. MMF has recently gained popu-
larity as a glucocorticoid- and calcineurin-inhibitor-sparing
agent for the treatment of children with idiopathic nephrotic
syndrome [18–21]. Lately, the first reports of its use in HSPN
have been published in adults [22–24] and children [25, 26].
However, pediatric experience is still limited.
Our goal was to assess the efficacy and safety of MMF in
HSPN presenting with nephrotic-range proteinuria and evalu-
ate the beneficial effect of early initiation of therapy adjusted
by therapeutic drug monitoring (TDM).
Methods
Patients
Our work is a single-center, retrospective study conducted at the
Division of Pediatric Nephrology, Children’s and Adolescents’
Hospital, University Hospital of Cologne, Germany. In the last
10 years (from December 2006 to December 2016), 18 children
with HSP having nephrotic-range proteinuria (defined as
>40 mg/m2BSA/h or 2 g/gCrea—in most cases determined
with a 24-h urine collection and in a few cases in spot urine
samples) despite renin–angiotensin–aldosterone system
(RAAS) blockade were biopsied in our center confirming the
diagnosis of HSPN (Fig. 1). The study was performed in accor-
dance with the Declaration of Helsinki 2013.
Histological analysis
Sixteen of 18 first renal biopsies were processed at the Renal
Pathology Laboratory, University Hospital of Cologne,
Germany, using standard techniques for light microscopy, im-
munofluorescence, and electron microscopy. Renal biopsies
were examined by the same pathologist (JUB) and classified in
accordance with the modified Oxford classification [27] and the
International Study of Kidney Disease in Children (ISKDC)
scoring system [28]. The two remaining biopsies (patients 11
and 16) had been processed elsewhere. On average, renal biopsy
was performed 54.0 ± 80.7 days after the first diagnosis of HSP.
Treatment protocol
Due to persisting proteinuria (either in nephrotic range or con-
stantly increasing) under RAAS blockade (ramipril in 14 pa-
tients, dual RAAS blockade with ramipril and losartan in four
Pediatr Nephrol

Fig. 1 Patient selection. This

retrospective, single-center study
paents with HSP
involved 18 patients presenting with proteinuria
with Henoch-Schönlein purpura
nephritis (HSPN) and nephrotic-
range proteinuria treated with RAAS inhibion
mycophenolate mofetil (MMF)
after insufficient therapeutic recovery proteinuria either in nephroc range or with increasing tendency
response to renin–angiotensin–

61.9 ± 95.4 days

aldosterone system (RAAS)
blockade and corticosteroids.
*One patient received paents with biopsy proven HSPN referred to our department for further therapy
corticosteroids orally, and seven or
paents referred to our department for biopsy and further treatment
received corticosteroids IV at a
dosage of 1–2 mg/kg per day
i.v./oral steroid therapy

85.0 ± 88.0 days

before referral to our department.
PU proteinuria, BSA body surface 18 paents
area recovery with biopsy proven HSPN
(mean se albumin: 34.5g/l;
mean PU: 113mg/m2BSA/h )

i.v. steroid pulse + oral steroids

proteinuria either in nephroc range or less than 50% decrease in

proteinuria

21.9 ± 17.5 days

18 paents with HSPN
(mean se albumin: 36.9g/l;
mean PU: 69mg/m2BSA/h)

MMF

2 responders
2 inial non-responders 14 stable responders relapsed aer
disconnuaon

children) and on the basis of histological confirmation of
HSPN, pulsed methylprednisolone (MP) was initiated on aver-
age 61.9 ± 95.4 days after the first diagnosis (300 mg/m2BSA/
day on 3 alternate days), followed by prednisone at 60 mg/
m2BSA/day for 3 weeks, then 40 mg/m2BSA/day on alternate
days for 4 weeks [29]. On average, 3 weeks (21.9 ± 17.5 days)
after initiation of pulsed MP, the course of proteinuria was re-
evaluated, and on the basis of this evaluation (<50% decrease in
proteinuria or proteinuria >40 mg/m2BSA/h or 2 g/gCrea) a
decision was made to start MMF treatment. Thus, MMF was
introduced as the second-line therapy on average 85.0 ± 88.
0 days after the first diagnosis of HSP, 33.8 ± 47.6 days after
the diagnosis of HSPN, and 21.9 ± 17.5 days after initiation of
pulsed MP. The initial dose of MMF was 800–1200 mg/
m2BSA/day divided into two daily doses, which was later ad-
justed according to the result of repeated TDM (performed
within 3 months of therapy onset and yearly thereafter). TDM
profiling was started 12 h after the last dose of MMF. For TDM,
we used a limited sampling strategy based on measurements at
0 (MPA-C0) 30 and 120 min after drug intake. Estimation of
individual estimated mycophenolic acid–area under the
concentration-time curve profiles (eMPA-AUC0–12h) was de-
rived from plasmatic MPA-concentrations and calculated based
on a two-compartment model using the pharmacokinetic soft-
ware package MW/Pharm (version 3.5).
Management of relapses and adverse events
Proteinuria was monitored at home by dipstick analysis and
24-h urine collection on regular follow-up visits. Partial remis-
sion was defined as proteinuria between 4 and 40 mg/m2BSA/
h or 0.2–2 g/gCrea and complete remission as proteinuria
<4 mg/m2BSA/h or 0.2 g/gCrea. Relapse was defined by pro-
teinuria >40 mg/m2BSA/h or 2 g/gCrea, or deterioration of
kidney function. A relapse was treated again with the thera-
peutic regimen described above. Adverse events, such as di-
arrhea (three or more watery stools per day), gastrointestinal
pain, weight loss, anemia, leukocytopenia, thrombocytopenia,

and infections were carefully monitored. In case of gastroin-
testinal toxicity not otherwise explained, hemoglobin levels
<8.0 g/dl, leukocyte counts <3.5 gpt/l, or thrombocyte count
<150 gpt/l, MMF therapy was discontinued.
Statistical analysis
Data were interpreted as mean and standard deviation (SD).
The relation between variables was examined with either
Pearson’s or Spearman’s correlation, depending on their dis-
tribution. Paired Student’s t tests or Wilcoxon signed-rank
tests were applied for data comparison likewise, depending
on their distribution. To find the cut-off level of eMPA-
AUC0–12h indicating sensitivity and specificity for prediction
of relapses, receiver operating characteristic (ROC) plots were
created. To compare blood results and clinical data over the
course of time, repeated measures analysis of variances
(ANOVA) were used, which were corrected by Greenhouse–
Geisser correction by violation of the assumption of spheric-
ity. Bonferroni-corrected t tests were performed as post hoc
tests. To assess other factors (time between diagnosis of HSP
and initiation of MMF therapy in days, Oxford total, Oxford
crescent, and ISKDC score) that have eventual influence on
early remission after MMF treatment, linear regression analy-
sis was performed. Results with p < 0.05 were considered sta-
tistically significant. IBM SPSS (version 24) and Microsoft
Excel were used for statistical analysis.
Observed individual MPA plasma concentrations were
fitted to an open two-compartment model employing a
Bayesian estimation method, taking into account observed
plasma levels and pharmacokinetic (PK) population parame-
ters. Using this two-stage PK analysis approach, an estimated
individual Bayesian posterior PK parameter (eMPA-AUC0-
12h) was calculated (US Department of Health and Human
Services, Food and Drug Administration. Guidance for
Industry on Population Pharmacokinetics, 1999). eMPA-
AUC0-12h values were targeted to be >30 mg*h/l [30].
Results
Table 1 represents the main clinical characteristics of 18 chil-
dren with HSPN who developed nephrotic-range proteinuria
and were treated with MMF as second-line therapy. Mean age
of our patients at disease onset was 9.6 ± 3.4 years, with a
strong male predominance (14 boys, 4 girls). The children were
followed up in our center for a mean of 42.2 ± 28.0 months.
Table 2 demonstrates the histological findings in 16 pa-
tients; 63% of biopsies showed advanced changes according
to ISKDC (ISKDC ≥ III), which were all accompanied by
crescent formation according to the modified Oxford
classification.
Pediatr Nephrol
After pulsed MP, a regular urine check was performed
weekly in the first month and thereafter every second week.
On average, 3 weeks after initiation of pulsed MP, ten of 18
patients still had nephrotic-range proteinuria (>40 mg/
m2BSA/h or 2 g/gCrea), and the remaining patients still had
mild to moderate proteinuria (4–40 mg/m2BSA/h or 0.2-2 g/
gCrea), with <50% decrease of initial proteinuria. Until this
time, we found no significant reduction in proteinuria (p =
0.053 or p = 0.151 in case of unit mg/m2BSA/h or g/gCrea,
respectively) or in serum albumin (p = 0.061). Therefore, pa-
tients were started on MMF after a mean of 85.0 ± 88.0 days
after presentation of HSP and 21.9 ± 17.5 days after initiation
of pulsed MP. Thereafter, the mean duration of MMF therapy
was 26.1 ± 11.5 months and mean follow-up after discontinu-
ation 32.9 ± 20.7 months. Two patients with initial ISKDC
score of IIIa/b and crescent formation did not improve with
MMF treatment. However, if a child improved (89%), no re-
lapse developed under MMF therapy. Two children experi-
enced a relapse after discontinuation (after 2 and 11 months,
respectively) (Table 1) and were treated again with the thera-
peutic regimen described above.
The most important effect of MMF treatment, a rapid and
significant decrease in proteinuria and increase in serum albu-
min, was seen after 1 month (Fig. 2, Table 3), without deteri-
oration of kidney function in short- or long-term follow-up
(Supplementary Table S1).
During MMF treatment, there was no significant change in
hemoglobin levels or thrombocyte counts (Supplementary
Table S1). However, a significant decline—in general, with-
out leaving the normal range—was observed in leukocyte
count between initiation of MMF therapy and the later time
points (months 1, 3, 12, last follow-up) (Supplementary
Table S1).
MPA daily dose and MPA-C0 correlated significantly with
the eMPA-AUC0-12h (p = 0.001), but rather weakly (r2 =
0.412 and r2 = 0.188, respectively).
Seeking the cause of nonresponse to MMF, we compared
the two nonresponders with the other 16 patients whose pro-
teinuria improved during therapy. However, no remarkable
differences in the following values were found during the
observation period: MMF dose was 1011 [95% confidence
interval (CI) 607–1416] mg/day vs. 916 (95% CI 840–992)
mg/day, MPA-C0 amounted to 2.5 (95% CI 0.8–4.2) mg/l vs.
2.8 (95% CI 2.0–3.7) mg/l and eMPA-AUC0-12h values were
49.0 (95% CI 26.0–72.0) mg*h/l vs. 52.6 (95% CI 44.2–61.1)
mg*h/l—nonresponders vs. responders, respectively.
To find a cut-off level of eMPA-AUC0-12 indicating sensi-
tivity and specificity for predicting complete remission at
month 3, year 1, and year 2, ROC curves were created. A
cut-off of 56.4 mg*h/l showed a sensitivity of 80% and a
specificity of 83.3% to discriminate children in complete re-
mission from the ones with proteinuria >4 mg/m2BSA/h
(ROC–AUC 0.833, p = 0.035) (Fig. 3a, Table 4a) and a
Pediatr Nephrol

Table 1 Main clinical characteristics of 18 children with Henoch-Schönlein purpura (HSP) who developed nephrotic-range proteinuria and were
treated with mycophenolate mofetil (MMF)

Age at Gender Time between HSP Duration of Total Follow-up after Favorable Relapse Relapse after
HSP onset and MMF MMF follow-up MMF response to during discontinuation
onset initialization (days) therapy (months) discontinuation MMF MMF of MMF therapy
(years) (months) (months) therapy therapy

1. 15.0 F 43 17.0 37.1 20.1 No NA NA

2. 8.4 M 43 18.0 61.0 43.0 No NA NA
3. 11.5 M 45 32.0 32.0 NA Yes No no
4. 5.9 M 76 NA 20.9 NA Yes No NA
5. 8.2 F 46 24.0 39.5 NA Yes No yes
6. 12.4 M 72 NA 16.3 NA Yes No NA
7. 10.0 M 78 25.4 40.1 14.7 Yes No No
8. 9.0 M 28 24.0 24.0 NA Yes No No
9. 4.9 F 9 19.8 64.5 44.7 Yes No No
10. 10.1 M 240 54.0 70.7 16.7 Yes No No
11. 5.5 F 36 23.9 23.9 NA Yes No No
12. 10.2 M 361 25.0 33.9 8.9 Yes No No
13. 14.0 M 120 8.0 21.1 NA Yes No Yes
14. 3.4 M 90 NA 6.2 NA Yes No NA
15. 15.4 M 47 15.0 15.0 NA Yes No No
16. 8.8 M 40 41.6 95.7 54.0 Yes No No
17. 8.7 M 46 26.5 50.8 24.4 Yes No No
18. 10.8 M 110 37.7 107.0 69.3 Yes No No
mean ± SD 9.6 ± 3.4 14 M; 85.0 ± 88.0 26.1 ± 11.5 42.2 ± 28.0 32.9 ± 20.7 88.9% 0% 15.4%
4F

MMF mycophenolate mofetil, M male; F female, SD standard deviation, NA not applicable

sensitivity of 75% and specificity of 83.3% to discriminate Discussion

children in complete remission from those with proteinuria
>0.2 g/gCrea (ROC–AUC 0.823, p = 0.017) (Fig. 3b,
Table 4b) 3 months after initiation of MMF. At later time
points (1 and 2 years after onset of MMF therapy), ROC curve
analysis failed to show an eMPA-AUC0-12 dependent advan-
tage (data not shown).
According to our experience, mild to moderate gastrointes-
tinal disturbances, mild anemia, and upper respiratory
infections were most frequently reported or measured during
MMF treatment. Comparing patients with eMPA-AUC0–12h
values below or above 56.4 mg*h/l, we found no relevant
differences in frequency of adverse events during the first
(most vulnerable) 3 months of treatment. Although diarrhea
was more frequent in patients with eMPA-AUC0-12h values
>56.4 mg*h/l, weight loss occurred with the same frequency
in both groups. Laboratory parameters did not differ
significantly (Supplementary Table S2). Importantly, diarrhea
regressed after dividing daily MMF into four doses, and
anemia resolved spontaneously with time. Other factors
(time between diagnosis of HSP and initiation of MMF
therapy in days, Oxford total, Oxford crescent, and ISKDC
score) did not influence early remission after MMF treatment
(data not shown).
Results of our retrospective, single-center study show that MMF
in combination with an initial course of corticosteroids may be
an effective agent to achieve and maintain remission in children
with HSPN and nephrotic-range proteinuria. Applying TDM
and adjusting eMPA-AUC0-12h to the value of 56.4 mg*h/l at
the onset of therapy, it is possible to achieve complete remission
already within 3 months without severe adverse effects.
Treatment options
HSPN in children was traditionally considered a self-limiting
disease requiring only supportive treatment. However, after
long-term follow-up studies revealed the high rate of late-
developing chronic kidney disease (CKD) in this population
[31], the need for effective anti-inflammatory treatment in sub-
groups is generally accepted [32]. CyA treatment in HSPN with
nephrotic-range proteinuria, however, has not found its way into
KDIGO guidelines [7], despite its promising effects [14, 15].
KDIGO guidelines suggest treatment with RAAS inhibitors for
3-6 months and addition of steroids orally for 6 months in pa-
tients with persisting nephrotic-range proteinuria. Only in case of
crescent formation in >50% of glomeruli is the use of
 Pediatr Nephrol

Table 2 Histological characteristic of 16 children with Henoch- HSPN accompanied by nephrotic-range proteinuria, pulsed MP
Schönlein purpura (HSP) who developed nephrotic-range proteinuria
therapy was administered directly after biopsy, resulting in slight
and were treated with mycophenolate mofetil (MMF). The table
demonstrates histological findings according to the Oxford improvements in proteinuria, but therapeutic response until re-
classification and International Study of Kidney Disease in Children evaluation was insufficient. However, we cannot formally rule
(ISKDC) at the first renal biopsy and long-term outcome out a late effect of pulsed MP on the course of the disease [10] or
Oxford ISKDC clinically assessed outcome a more favorable therapeutic effect of pulsed MP at a higher
dose, as others recommend [10, 15]. MMF was hence introduced
1. M1;E1;S1;T2;C2 IIIb Nonresponder as early as 3 months (85 ± 88.0 days) after disease onset to re-
2. M0;E1;S1;T0;C1 IIIa Nonresponder duce disease activity (endocapillary inflammation and crescent
3. M0;E0;S1;T0;C2 IVa good formation). However, this critical time frame may be further
4. M0;E1;S0;T0;C1 III good reduced by 4–6 weeks, as some authors suggest [2, 10, 12, 15,
5. M1;E1;S1;T0;C2 IIIb relapse after MMF discontinuation 25, 26, 34], if future referrals of patients are carried out resulting
6. M1;E1;S1;T0;C2 IIIb good in timely diagnostic kidney biopsy and consecutive treatment
7. M0;E1;S0;T0;C0 II good without delay.
8. M1;E1;S0;T0;C0 II good
9. M1;E1;S1;T0;C0 II good Mechanism of action
10. M0;E0;S0;T0;C0 II good
12. M0;E0;S0;T0;C0 II good The rationale for using MMF therapy in HSPN patients is the
13. M0;E0;S0;T0;C1 IIIa relapse after MMF discontinuation well-known mechanism of its active moiety, MPA, which in-
14. M0;E1;S1;T0;C1 III good hibits lymphocyte inosine-5′-monophosphate dehydrogenase
15. M1;E0;S1;T0;C0 II good (IMPDH). Thereby, it strongly decreases guanosine-5′-triphos-
17. M0;E0;S1;T0;C2 III good phate (GTP) concentrations essential for lymphocyte prolifera-
18. M0;E0;S1;T0;C1 III good tion [17]. In HSPN, MPA can decrease IgA production of lym-
phocytes and reverse even the IgA1-aberrant O-glycosylation
M mesangial hypercellularity, E endocapillary hypercellularity, S segmen-
tal glomerulosclerosis, T tubular atrophy/interstitial fibrosis, C crescent
level in peripheral lymphocytes [35], resulting in reduced de-
formation position of subendothelial and mesangial IgA-containing im-
mune complexes. It is less known, but also important in the
cyclophosphamide a weak suggestion. This guideline refers to pathomechanism of mesangial proliferation, that MPA also sig-
studies conducted in adult patients with IgA nephritis, which is a nificantly depletes GTP levels in monocytes [17]. Beyond this
much more chronic, indolent disease, with continuous mesangial classical effect, a vast body of evidence has shown that this
proliferation in contrast to HSPN, with a high frequency of cres- agent effects not only immune cells but also nonimmune cells
cent formation, endocapillary infiltration of neutrophils, and sub- [36]. MPA exerts a direct inhibitory effect on secretion of osteo-
sequent acute and deteriorating episodes [33]. We therefore pontin, fibronectin, and collagen type I in stimulated mesangial
followed a stricter treatment protocol, which aims to reduce de- cell cultures in vitro [37, 38] and on production of collagen type
lay in IS treatment initiation and prevent undertreatment of cres- III in mesangial cells in vivo after 5/6 nephrectomy [39].
centic lesions. Accordingly, in our patients with biopsy-proven Furthermore, Takeda et al. showed that in an anti-glomerular-
basement-membrane (GBM) antibody-induced rat model, 14-
day MMF treatment prevented glomerular crescent formation
50 120
and glomerulosclerosis if MMF treatment was started on day 1,
proteinuria (mg/m2BSA/h)

47
100 but not when started 5 days after the anti-GBM antibody injec-
tion [40].
se albumin (g/l)

80
44
60
41
40 Clinical data and laboratory values
38
20
In the literature, 15 children with HSPN treated with MMF due
35 0 to steroid-resistant nephrotic-range proteinuria have been de-
scribed so far (Table 5). It is well known that nephrotic-range
proteinuria lasting >3 months has a significant adverse effect on
renal outcome [31]. Accordingly, in the children mentioned
Fig. 2 Changes in serum albumin (gray) and proteinuria (black) in the course
of disease. To compare parameters in the first year, repeated measures
above as well as our patients, the decision to initiate MMF
analysis of variance (ANOVA) was performed. BSA body surface area. therapy was made within 3 months after disease onset to reduce
* P < 0.05 vs. initiation of mycophenolate mofetil (MMF) therapy. disease activity and improve long-term renal outcome. In the
Pediatr Nephrol
Table 3 Changes in serum albumin and proteinuria in the course of disease. To compare parameters, in the first year, repeated measures analysis of
variance (ANOVA) was performed
at start of MMF month 1
Serum albumin (g/l) 36.9 ± 5.1 40.0 ± 4.0 *
proteinuria (mg/m2BSA/h) 69.0 ± 46.5 24.9 ± 28.9 *
proteinuria (g/gCrea) 2.8 ± 2.3 1.2 ± 1.4 *
BSA body surface area, MMF mycophenolate mofetil
*P < 0.05 vs. initiation of MMF therapy
two pediatric studies [25, 26] listed in Table 5, MMF therapy
was administered for either 4 and 12 months, respectively, at a
dosage of 20–30 mg/kg per day. Both studies achieved a favor-
able response to MMF in 100% of cases and found no relapse
during or after MMF treatment. However, the small number of
patients is certainly a limitation of both studies. In contrast to
these reports, our patients were treated with MMF for a longer
period—at least 6 months of complete remission (proteinuria
<4 mg/m2BSA/h or 0.2 g/gCrea, no efflorescences, no joint
pain) before discontinuing MMF therapy. The higher ratio of
non-responders (11.1%) as well as relapses after MMF discon-
tinuation (15.4%) in our study may be explained by the larger
study population than in the two previous reports. Comparison
of our results with data of adult patients with HSPN is difficult,
since adult nephrologists initiate MMF therapy later than pedi-
atric nephrologists. Nevertheless, the higher patient number in
the adult studies gives us a more realistic picture of the percent-
age of non-responders and relapse rate during and after MMF
treatment [23, 24], which were comparable with our
experiences.
A B
Sensivity
Sensivity
AUC=0.833
p=0.035
1-Specificity
Fig. 3 Receiver operating characteristics (ROC) plot calculated for
estimated mycophenolic acid area under the curve (eMPA–AUC0-12h)
3 months after initiation of mycophenolate mofetil (MMF) therapy. a
Discrimination of children in complete remission from children with
month 3 month 6 month 12
41.9 ± 3.7 * 43.9 ± 4.2 * 44.4 ± 3.8 *
18.6 ± 17.7 * 12.4 ± 10.8 * 10.2 ± 11.9 *
0.8 ± 0.8 * 0.5 ± 0.4 * 0.4 ± 0.5 *
Our patients showed a significant decrease in proteinuria
and increase in serum albumin as early as 1 month after initi-
ation of therapy, which stayed in normal range for the rest of
the follow-up (42.2 ± 28.0 months). However, only one pa-
tient reached complete remission within the first month and
eight patients within 3 months. The time course of proteinuria
reduction in response to MMF treatment is in accordance with
the limited published data available. As early as 1 month of
MMF treatment initiation, one can expect a significant im-
provement in proteinuria [23–25]. However, achieving com-
plete remission (< 4 mg/m2BSA/h or 0.2 g/gCrea) takes more
time: 15.4 ± 17.2 months in our patients, 6.1 ± 3.7 months in
the cohort described by Ren et al. [23], and 11 ± 1.7 months in
the cohort described by Du et al. [25]. We observed a signif-
icant decline in leukocyte count after initiating MMF therapy,
with no overt risk of leukocytopenia. It is highly possible that
leukocyte counts of our patients were stimulated by the pre-
ceding use of corticosteroids that diminished with time. All
other monitored parameters were in the physiological range
and did not change relevantly within the observation period.
AUC=0.823
p=0.017
1-Specificity
proteinuria >4 mg/m2 body surface area (BSA)/h (ROC-AUC: 0.833,
p = 0.035). b Discrimination of children in complete remission from
those with proteinuria >0.2 g/gCrea (ROC–AUC: 0.823, p = 0.017).
 Pediatr Nephrol

Table 4 Receiver operating characteristics (ROC) plot calculated for with nephrotic syndrome [20, 21, 43], we used this parameter
estimated mycophenolic acid area under the curve (eMPA-AUC0-12h)
for monitoring, which necessitated a dose adjustment in 50%
3 months after initiation of mycophenolate mofetil (MMF) therapy. A
cut-off of 56.4 mg*h/l showed a sensitivity of 80% and a specificity of of patients. ROC curve analysis showed that it is essential in
83.3% to discriminate children in complete remission from children with the first 3 months to reach an eMPA-AUC0-12h close to or
proteinuria >4 mg/m2 body surface area (BSA)/h (p = 0.035) and a >56.4 mg*h/l, since complete remission can be achieved with
sensitivity of 75% and specificity of 83.3% to discriminate children in
high sensitivity (80%) and specificity (83.3%) within
complete remission from those with proteinuria >0.2 g/gCrea (p = 0.017)
3 months. This eMPA-AUC0–12h is at the upper limit of the
eMPA-AUC (mg*h/l) Sensitivity 1-Specificity target range recommended in immunosuppressive combina-
tion therapy after kidney transplantation (30–60 mg*h/l [30])
ROC curve analysis for complete remission
(proteinuria <4 mg/m2BSA/h) at month 3
and in line with the suggestion of higher target ranges in chil-
32.2 1.000 0.667
dren with idiopathic nephrotic syndrome on MMF monother-
apy (> 50 mg*h/l [20], > 45 mg*h/l [21, 43]). At later time
38.5 0.800 0.667
points (at the end of the first and second year of treatment), we
44.4 0.800 0.583
found no cut-off level for eMPA-AUC0-12h above which
49.1 0.800 0.500
MMF therapy would be more effective. Therefore, we hypoth-
53.5 0.800 0.417
esize that a lower eMPA-AUC0-12h target is sufficient later in
54.7 0.800 0.333
the course of treatment. This phenomenon might be explained
54.9 0.800 0.250
by the pathomechanism of HSPN; in the first acute phase of
56.4 0.800 0.167
disease, the inflammatory component is very pronounced,
59.9 0.600 0.167
while it subsides later on. Therefore, our result seems to be
64.3 0.600 0.083
in line with pathophysiological processes and emphasizes the
71.1 0.600 0.000
importance of early anti-inflammatory treatment.
ROC curve analysis for complete remission
(proteinuria <0.2 g/gCrea) at month 3
32.2 1.000 0.583
Adverse events
38.5 0.875 0.583
In order to analyze whether the children with an eMPA-
44.4 0.750 0.583
AUC0-12h above 56.4 mg*h/l develop adverse events more
49.1 0.750 0.500
frequently than those with an eMPA-AUC 0-12h below
53.5 0.750 0.417
56.4 mg*h/l, the two groups were compared. Although the
54.7 0.750 0.333
patients with higher exposure tended to experience diarrhea
54.9 0.750 0.250
more frequently, it was ameliorated after splitting the daily
56.4 0.750 0.167
MMF amount into four doses. This finding is in line with
59.9 0.625 0.167
the literature showing no associations between adverse events
63.1 0.625 0.083
and systemic concentration of total MPA in pediatric renal
71.1 0.500 0.000
transplant patients [44, 45]. An association was found only
between adverse events and the free (not albumin-bound)
MPA [41, 45], which was not determined in our study.
This finding is in line with the reported favorable profile of
MMF regarding potential side effects. Limitations

Therapeutic drug monitoring
The important role of TDM during MMF treatment has been
demonstrated after renal transplantation and in patients with
nephrotic-range proteinuria. Measuring eMPA-AUC can re-
duce the risk of acute rejection in renal allograft recipients
[41, 42]. Additionally, monitoring eMPA-AUC can prevent
relapses in pediatric patients with frequently-relapsing ne-
phrotic syndrome [20, 21, 43]. To further optimize our treat-
ment protocol, TDM was applied for the first time—to the best
of our knowledge - to treat HSPN with MMF. Considering
that eMPA-AUC0-12h reflects total body drug exposure and
shows the strongest correlation with relapses in other diseases
There are limitations to this retrospective study. It does not
exclusively prove the beneficial therapeutic effect of MMF,
because a late effect of pulsed MP on disease course cannot be
ruled out. Furthermore, our children constituted the largest
published pediatric population with HSPN treated with
MMF. However, this patient number still must be considered
small, and no control group was available. HSPN is a rare
disease. Therefore, multicentric, prospective, randomized,
TDM-based studies with larger patient numbers are difficult
to implement but are highly desirable to meet the standards of
evidence-based medicine. In particular, the proposed eMPA-
AUC 0-12h cutoff needs to be confirmed and validated
prospectively.
 Pediatr Nephrol

Table 5 Summary of cases in the available literature on Henoch-Schönlein purpura nephritis treated with mycophenolate mofetil (MMF)

number of total control arm MMF arm time between disease duration initial dosage MMF favorable relapse relapse follow-up
patients follow-up onset and MMF of MMF of MMF adjusted response to during MMF after MMF after MMF
on MMF (months) initialization (days) therapy according MMF therapy (%) discontinuation discontinuation
(mo) to TDM therapy (%) (%) (months)

Dede et al. adult 1 42 none MMF, oral NI 24 2000 mg/day no 100 no 0 18

2008 steroids
[22]
Ren et al. 27 28 oral steroids MMF, oral NI NI 1000 mg/day no 78 no 0 NI
2012 steroids
[23]
Han et al. 33 28 oral steroids MMF, oral 135 12 1000 mg/day no 73 no 6 NI
2015 steroids
[24]
Du et al. pediatric 12 47 none MMF, MP 28 12 20–25 mg/kg/day no 100 no 0 35
2012 pulse, oral
[25] steroids
Nikibakhsh 3 8 none MMF, MP 24 4 30 mg/kg day no 100 no 0 5
et al. pulse, oral
2014 steroids
[26]
our patients 18 42 none MMF, MP 82 26 800–1200 mg/BSA yes 89 no 15 33
pulse, oral
steroids

NI no information, MP methylprednisolone, TDM therapeutic drug monitoring, BSA body surface area

Conclusion
Our retrospective single-center study demonstrates that fol-
lowing initial administration of corticosteroids, MMF treat-
ment may represent a safe and potentially effective secondary
treatment option for children with HSPN accompanied by
nephrotic-range proteinuria. Furthermore, we propose a time-
ly diagnosis and early initiation of MMF therapy with an
eMPA-AUC0–12h target value of 56.4 mg*h/l to achieve com-
plete remission within 3 months to resolve severe inflamma-
tory glomerular lesions and avoid progression to CKD.
Compliance with ethical standards
Conflict of interest The authors declare no conflict of interest.
References
1. Aalberse J, Dolman K, Ramnath G, Pereira RR, Davin JC (2007)
Henoch Schonlein purpura in children: an epidemiological study
among Dutch paediatricians on incidence and diagnostic criteria.
Ann Rheum Dis 66:1648–1650
2. Davin JC, Coppo R (2014) Henoch-Schönlein purpura nephritis in
children. Nat Rev Nephrol 10:563–573
3. Gardner-Medwin JM, Dolezalova P, Cummins C, Southwood TR
(2002) Incidence of Henoch-Schonlein purpura, Kawasaki disease,
and rare vasculitides in children of different ethnic origins. Lancet
360:1197–1202
4. Pohl M (2015) Henoch-Schönlein purpura nephritis. Pediatr
Nephrol 30:245–252
5. Jauhola O, Ronkainen J, Koskimies O, Ala-Houhala M, Arikoski P,
Hölttä T, Jahnukainen T, Rajantie J, Ormälä T, Turtinen J, Nuutinen
M (2010) Renal manifestations of Henoch-Schonlein purpura in a
6-month prospective study of 223 children. Arch Dis Child 95:877–
882
6. North American Pediatric Renal Trials and Collaborative Studies;
Annual Transplant Report (2014). Available from: https://web.
emmes.com/study/ped/annlrept/annualrept2014.pdf
7. KDIGO Work Group (2012) KDIGO clinical practice guideline for
glomerulonephritis, Chapter 11: Henoch-Schönlein purpura nephri-
tis. Kidney Int Suppl 2, 218–220. https://doi.org/10.1038/kisup.
2012.24
8. Hahn D, Hodson EM, Willis NS, Craig JC (2015) Interventions for
preventing and treating kidney disease in Henoch-Schönlein
Purpura (HSP). Cochrane Database Syst Rev 7(8):CD005128
9. Weiss PF, Feinstein JA, Luan X, Burnham JM, Feudtner C (2007)
Effects of corticosteroid on Henoch-Schönlein purpura: a system-
atic review. Pediatrics 120:1079–1087
10. Niaudet P, Habib R (1998) Methylprednisolone pulse therapy in the
treatment of severe forms of Schönlein-Henoch purpura nephritis.
Pediatr Nephrol 12:238–243
11. Li Volti S, Li Volti G (2003) Early treatment with oral immunosup-
pressants in severe proteinuric purpura nephritis. Pediatr Nephrol
18:1197–1198
12. Tanaka H, Suzuki K, Nakahata T, Ito E, Waga S (2003) Early
treatment with oral immunosuppressants in severe proteinuric pur-
pura nephritis. Pediatr Nephrol 18:347–350
13. Tarshish P, Bernstein J, Edelmann CM Jr (2004) Henoch-Schönlein
purpura nephritis: course of disease and efficacy of cyclophospha-
mide. Pediatr Nephrol 19:51–56
Pediatr Nephrol
14. Park JM, Won SC, Shin JI, Yim H, Pai KS (2011) Cyclosporin A
therapy for Henoch-Schönlein nephritis with nephrotic-range pro-
teinuria. Pediatr Nephrol 26:411–417
15. Jauhola O, Ronkainen J, Autio-Harmainen H, Koskimies O, Ala-
Houhala M, Arikoski P, Hölttä T, Jahnukainen T, Rajantie J, Ormälä
T, Nuutinen M (2011) Cyclosporine a vs. methylprednisolone for
Henoch-Schönlein nephritis: a randomized trial. Pediatr Nephrol
26:2159–2166
16. Kawasaki Y, Suyama K, Hashimoto K, Hosoya M (2011)
Methylprednisolone pulse plus mizoribine in children with
Henoch-Schoenlein purpura nephritis. Clin Rheumatol 30(4):
529–535
17. Allison AC, Eugui EM (2000) Mycophenolate mofetil and its
mechanisms of action. Immunopharmacology 47:85–118
18. Bagga A, Hari P, Moudgil A, Jordan SC (2003) Mycophenolate
mofetil and prednisolone therapy in children with steroid-
dependent nephrotic syndrome. Am J Kidney Dis 42:1114–1120
19. Baudouin V, Alberti C, Lapeyraque AL, Bensman A, André JL,
Broux F, Cailliez M, Decramer S, Niaudet P, Deschênes G, Jacqz-
Aigrain E, Loirat C (2012) Mycophenolate mofetil for steroid-
dependent nephrotic syndrome: a phase II Bayesian trial. Pediatr
Nephrol 27:389–396
20. Gellermann J, Weber L, Pape L, Tönshoff B, Hoyer P, Querfeld U,
Gesellschaft für Pädiatrische Nephrologie (GPN) (2013)
Mycophenolate mofetil versus cyclosporin a in children with fre-
quently relapsing nephrotic syndrome. J Am Soc Nephrol 24:1689–
1697
21. Hackl Á, Cseprekál O, Gessner M, Liebau MC, Habbig S, Ehren R,
Müller C, Taylan C, Dötsch J, Weber LT (2016) Mycophenolate
Mofetil therapy in children with Idiopathic Nephrotic Syndrome:
does therapeutic drug monitoring make a difference? Ther Drug
Monit 38:274–279
22. Dede F, Onec B, Ayli D, Gonul II, Onec K (2008) Mycophenolate
mofetil treatment of crescentic Henoch-Schönlein nephritis with
IgA depositions. Scand J Urol Nephrol 42:178–180
23. Ren P, Han F, Chen L, Xu Y, Wang Y, Chen J (2012) The combi-
nation of mycophenolate mofetil with corticosteroids induces re-
mission of Henoch-Schönlein purpura nephritis. Am J Nephrol
36:271–277
24. Han F, Chen LL, Ren PP, Le JY, Choong PJ, Wang HJ, Xu Y, Chen
JH (2015) Mycophenolate mofetil plus prednisone for inducing
remission of Henoch-Schönlein purpura nephritis: a retrospective
study. J Zhejiang Univ Sci B 16:772–779
25. Du Y, Hou L, Zhao C, Han M, Wu Y (2012) Treatment of children
with Henoch-Schönlein purpura nephritis with mycophenolate mo-
fetil. Pediatr Nephrol 27:765–771
26. Nikibakhsh AA, Mahmoodzadeh H, Karamyyar M, Hejazi S,
Noroozi M, Macooie AA (2014) Treatment of severe henoch-
schonlein purpura nephritis with mycophenolate mofetil. Saudi J
Kidney Dis Transpl 25:858–863
27. Trimarchi H, Barratt J, Cattran DC, Cook HT, Coppo R, Haas M,
Liu ZH, Roberts IS, Yuzawa Y, Zhang H, Feehally J (2017) IgAN
Classification Working Group of the International IgA
Nephropathy Network and the Renal Pathology Society;
Conference Participants. Oxford classification of IgA nephropathy
2016: an update from the IgA nephropathy classification working
group. Kidney Int 91:1014–1021
28. Heaton JM, Turner DR, Cameron JS (1977) Localization of glo-
merular Bdeposits^ in Henoch–Schönlein nephritis. Histopathology
1:93–104
29. Pohl M, Dittrich K, Ehrich JHH, Hoppe B, Kemper MJ, Klaus G,
Schmitt CP, Hoyer PF (2013) Gesellschaft für Pädiatrische
Nephrologie (GPN) (2013) Behandlung der Purpura-Schönlein-
Henoch-Nephritis bei Kindern und Jugendlichen;
Therapieempfehlungen der Gesellschaft für Pädiatrische
Nephrologie (GPN). Monatsschr Kinderheilkd 161:543–553
Pediatr Nephrol
30. Weber LT, Hoecker B, Armstrong VW, Oellerich M, Tönshoff B
(2008) Long-term pharmacokinetics of mycophenolic acid in pedi-
atric renal transplant recipients over 3 years posttransplant. Ther
Drug Monit 30:570–575
31. Wakaki H, Ishikura K, Hataya H, Hamasaki Y, Sakai T, Yata N,
Kaneko T, Honda M (2011) Henoch-Schönlein purpura nephritis
with nephrotic state in children: predictors of poor outcomes.
Pediatr Nephrol 26:921–925
32. Coppo R, Amore A (2004) New perspectives in treatment of glo-
merulonephritis. Pediatr Nephrol 19:256–265
33. Davin JC, Coppo R (2013) Pitfalls in recommending evidence-
based guidelines for a protean disease like Henoch-Schönlein pur-
pura nephritis. Pediatr Nephrol 28:1897–1903
34. Ronkainen J, Ala-Houhala M, Huttunen NP, Jahnukainen T,
Koskimies O, Ormälä T, Nuutinen M (2003) Outcome of
Henoch-Schoenlein nephritis with nephrotic-range proteinuria.
Clin Nephrol 60:80–84
35. Xie L, Tan C, Fan J, Fu P, Tang Y, Tao Y, Qin W (2013)
Mycophenolic acid reverses IgA1 aberrant glycosylation through
up-regulating Cosmc expression in IgA nephropathy. Int Urol
Nephrol 45:571–579
36. Hackl A, Ehren R, Weber LT (2016) Effect of mycophenolic acid in
experimental, nontransplant glomerular diseases: new mechanisms
beyond immune cells. Pediatr Nephrol 32:1315–1322
37. Wang W, Mo S, Chan L (1999) Mycophenolic acid inhibits PDGF-
induced osteopontin expression in rat mesangial cells. Transplant
Proc 31:1176–1177
38. Dubus I, Vendrely B, Christophe I, Labouyrie JP, Delmas Y, Bonnet
J, Combe C (2002) Mycophenolic acid antagonizes the activation
of cultured human mesangial cells. Kidney Int 62:857–867
39. Badid C, Vincent M, McGregor B, Melin M, Hadj-Aissa A,
Veysseyre C, Hartmann DJ, Desmouliere A, Laville M (2000)
Mycophenolate mofetil reduces myofibroblast infiltration and col-
lagen III deposition in rat remnant kidney. Kidney Int 58:51–61
40. Takeda S, Takahashi M, Sado Y, Takeuchi K, Hakamata Y, Shimizu
H, Kaneko T, Yamamoto H, Ito C, Ookawara S, Asano Y, Kusano
E, Kobayashi E (2004) Prevention of glomerular crescent formation
in glomerulonephritis by mycophenolate mofetil in rats. Nephrol
Dial Transplant 19:2228–2236
41. Weber LT, Shipkova M, Armstrong VW, Wagner N, Schütz E,
Mehls O, Zimmerhackl LB, Oellerich M, Tönshoff B (2002) The
pharmacokinetic pharmacodynamic relationship for total and free
mycophenolic acid in pediatric renal transplant recipients: a report
of the german study group on mycophenolate mofetil therapy. J Am
Soc Nephrol 13:759–768
42. Le Meur Y, Büchler M, Thierry A, Caillard S, Villemain F, Lavaud
S, Etienne I, Westeel PF, Hurault de Ligny B, Rostaing L, Thervet
E, Szelag JC, Rérolle JP, Rousseau A, Touchard G, Marquet P
(2007) Individualized mycophenolate mofetil dosing based on drug
exposure significantly improves patient outcomes after renal trans-
plantation. Am J Transplant 7:2496–2503
43. Tellier S, Dallocchio A, Guigonis V, Saint-Marcoux F, Llanas B,
Ichay L, Bandin F, Godron A, Morin D, Brochard K, Gandia P,
Bouchet S, Marquet P, Decramer S, Harambat J (2016)
Mycophenolic acid pharmacokinetics and relapse in children with
steroid-dependent idiopathic Nephrotic syndrome. Clin J Am Soc
Nephrol 11:1777–1782
44. Hale MD, Nicholls AJ, Bullingham RE, Hené R, Hoitsma A,
Squifflet JP, Weimar W, Vanrenterghem Y, Van de Woude FJ,
Verpooten GA (1998) The pharmacokinetic-pharmacodynamic re-
lationship for mycophenolate mofetil in renal transplantation. Clin
Pharmacol Ther 64:672–683
45. Oellerich M, Shipkova M, Schütz E, Wieland E, Weber L, Tönshoff
B, Armstrong VW (2000) Pharmacokinetic and metabolic investi-
gations of mycophenolic acid in pediatric patients after renal trans-
plantation: implications for therapeutic drug monitoring. German
Study Group on Mycophenolate Mofetil Therapy in Pediatric
Renal Transplant Recipients. Ther Drug Monit 22:20–26