Bilirubine: een veelkoppig monster ?
SKML Sectie Algemene Chemie Gebruikersdag 3 juni 2010 prof dr ir Huib L. Vader
Bilirubin fractionation by HPLC
Delta bilirubine • In de circulatie gevormd uit geconjugeerd bilirubine en albumine • Langzame, niet-enzymatische vervanging van glucuronzuur door albumine (als GHb) • Pas aantoonbaar na een langdurige periode van geconjugeerde hyperbilirubinemie • Verdwijnt langzaam uit de circulatie. T1/2 als albumine 2-3 weken
Maisels M, McDonagh A. N Engl J Med 2008;358:920-928
Do we measure bilirubin correctly anno 2005?
•
Observaties in hoge concentratiegebied (300-400 µmol/l)
•
30% verschil tussen vloeistof chemie (Merck Mega en Roche Modular) en droge chemie (Vitros) – Mega en Modular 20% verhoogd t.o.v. ref. methode – Vitros 10% verlaagd t.o.v. referentiemethode
•
SRM916 wordt in beide systemen zeer goed terug gevonden
•
Discrepanties lijken te worden veroorzaakt door de aanwezigheid van geconjugeerd of δ-bilirubine
Relatie Vitros 250 Modular in monsters volwassenen
Bilirubin Fractions by Method
HPLC Peak Bilirubin Species
α Unconjugated
γ
δ
Singly
Doubly
Conjugated to
Conjugated
Conjugated
Albumin
Total
Traditional Methods
β
Indirect (Total – Direct)
Direct
?
Total Vitros Methods
Unconjugated (Bu)
Conjugated (Bc) Delta (Total –
Neonatal (Bu+Bc)
[Bu+Bc]) Direct (Total – Bu)
Correlatie tussen Modular/Vitros en referentiemethode
Referentie-techniek (totaal bilirubine cf. Jendrassik en Gróf
1. Cafeïnereagens (accelerator) wordt toegevoegd aan monster 2. Toevoeging gediazoteerd sulfanylzuur 3. Geconjugeerd en ongeconjugeerd reageren tot azobilirubine (rose) 4. Alkalische tartraatoplossing voor pH shift 5. Meting blauw/groene kleur bij 600nm
SRM916 gemeten op Modular en Vitros In relatie tot meting met referentiemethode
Guidelines for Intensive Phototherapy in Hospitalized Infants Born at a Gestational Age of 35 Weeks or More
Maisels M, McDonagh A. N Engl J Med 2008;358:920-928
Correlatie Vitros/Modular in neonatale monsters
Interlaboratory variability in albumin and bilirubin measurements on neonatal intensive care units. Christian V. Hulzebos, Peter H. Dijk, Deirdre E. van Imhoff, Arend F. Bos, on behalf of the BARTrial Studygroup Dept. of Paediatrics, Div. of Neonatology, University Medical Center Groningen, the Netherlands
Background Guidelines are used for the management of unconjugated hyperbilirubinemia to standardize care of jaundiced newborn infants. Treatment of unconjugated hyperbilirubinemia is based on total serum bilirubin. Measurement of albumin is recommended because low albumin is considered a risk factor for bilirubin encephalopathy. Variability in measurements of bilirubin and albumin levels may affect timing of starting or stopping treatment. To date, it is unclear whether variability exists in bilirubin or albumin measurements on laboratories of Dutch neonatal intensive care units (NICUs).
Bilirubin levels of the QC samples were overestimated and a large interlaboratory variation was found (Figure 1A & B).
Figure 1. Measured Bilirubin Levels vs. Bilirubin Levels of Quality Control Samples for Albumin Levels of 10 (Fig.1A) and 20 g/L (Fig.1B). Fig. 1A. Albumin = 10 g/L. 600
500
500
Bilirubin 400 measured (µmol/L) 300
400
200
200
100
100
300
0
Objective We aimed to assess the variability in measurements of albumin and bilirubin on Dutch NICUs. Design/Methods Stabilized quality control samples with different levels of albumin (0, 10, 15, 20, 25 and 30 g/L) and bilirubin (100, 200, 300, 400, and 500 µmol/L) were sent to laboratories of all Dutch NICUs (n=10). The mean, standard deviation (SD) and coefficients of variations (CV) were calculated per sample.
Fig. 1B. Albumin = 20 g/L.
600
0 0
100 200 300 400 500 600
0
100 200 300 400 500 600
Bilirubin levels (µmol/L) in Quality Control Samples
Differences in methodology may contribute to the observed interlaboratory variability. However, comparison of bilirubin measurements of laboratories using the same device yielded similar accuracy and variabilty (Figure 2).
Figure 2. Measured Bilirubin Levels vs. Bilirubin Levels of Quality Control Samples using the same method (Roche Modular (2A). Fig. 2A. Roche Modular 600 500
Bilirubin 400 measured 300 (µmol/L) 200 100 0
Combinations of Albumin and Bilirubin levels in Quality Control Samples. Values in the blue boxes are used in Figures 1A and 1B.
Results Measured albumin levels were ~ 10% lower than levels of albumin in quality control samples. Maximal CV was 6.8 % (Table). Maximal CV of measured bilirubin levels was 9.2% (data not shown). Table. Results of measurements of albumin levels.
Albumin (g/L) 30 25 20 15 10
Mean ± SD 27.2 ± 1.0 22.6 ± 0.9 18 ± 0.7 13.3 ± 0.7 8.9 ± 0.6
CV 3.5 4.0 3.7 5.3 6.8
Data represent mean ± SD, or coefficients of variations (CV=SD/ mean)
0
100 200 300 400 500 600
Bilirubin levels (µmol/L) in Quality Control Samples
Conclusions 1. Accuracy of albumin and bilirubin measurements on laboratories of Dutch NICUs is poor. 2. Considerable variability in albumin and bilirubin measurements exists between laboratories of Dutch NICUs potentially affecting treatment of jaundiced newborn infants. Recommendations to improve care of jaundiced newborn infants 1. Accuracy of albumin and bilirubin measurements needs to be improved by recalibration. 2. Quality Control Samples with age-specific ranges of albumin and bilirubin should be used to reduce the large interlaboratory variability.
[email protected]
Na correctie voor onzuiverheid in SRM916a varieerden de resultaten in monster 5 van 344 tot 346 µmol/l
2009-2
2009-10
34
34
Roche
Ref C
32
Ortho Vitros
32
30
30
28
28
26
26
24
24
22
Roche labs
22
Vitros labs
20
Target Reference labs
20
Target Referecence lab Mean Vitros labs
Mean Roche labs
18
18
Measured by Manufacturer
Measured by Manufacturer
16
16 55
60
65
70
75
80
85
90
55
95
60
65
70
75
80
85
90
95
Ref A
34
34
Beckman 32
Siemens Dimension
32
30
30
28
28
26
26
24
24
Beckman labs
22
22
Dimension labs
Mean Beckman labs
20
Target Reference lab
Measured by Manufacturer
18
Target Reference lab
20 18
Mean Dimension labs
16 55
60
65
70
75
80
85
90
95
34
Measured by Manufacturer
16 55
60
65
70
75
80
85
90
95
34
Abbott
32
Siemens ADVIA
32
30
30
28
28
26
26
24
24
22
Abbott labs
20
Target Reference lab
Target Reference lab
20
Mean Abbott labs 18
ADVIA labs
22
Mean ADVIA labs
18
Measured by Manufacturer
16
Measured by Manufacturer
16 55
60
65
70
75
80
85
90
95
55
60
65
70
75
80
85
90
95
Number of participants per method, mean and SD in February 2009, September 2009 and March 2010
Methode
Number of Labs
Low Level
High level
Concentration
SD
Concentration
SD
Feb 2009
Sept 2009
Mrt 2010
Feb 2009
Sept 2009
Mrt 2010
Feb 2009
Sept 2009
Mrt 2010
Feb 2009
Sept 2009
Mrt 2010
Feb 2009
Sept 2009
Mrt 2010
Abbott
19
20
19
28.0
28.4
21.1
2.6
2.5
1.3
75.7
72.5
76.8
8.7
6.4
4.6
Siemens ADVIA
7
4
3
29.4
24.5
21.5
1.2
3.3
0.7
76.0
75.1
78.9
2.6
2.2
0.2
Siemens Dimension
7
11
17
27.1
25.8
19.5
1.0
2.0
1.4
70.2
71.6
76.2
1.3
2.7
3.9
Beckman
42
42
41
29.9
30.3
23.0
2.2
2.0
2.0
74.3
73.9
79.0
2.6
3.2
3.0
Olympus
-
-
9
-
-
22.6
-
-
2.1
-
-
78.1
-
-
2.9
Roche
99
117
120
27.1
26.4
19.2
2.4
2.2
1.9
73.1
70.6
73.0
5.6
5.6
5.8
Roche without high
-
-
103
-
-
18.8
-
-
1.6
-
-
71.6
-
-
3.8
Ortho Vitros
9
3
3
20.6
20.6
17.2
1.2
2.4
1.6
59.1
53.6
63.1
2.1
1.0
6.3
26.7
26.7
20.0
68.7
68.7
72.6
Target
Table 2. Measured Bilirubin levels as percentage of the value of the reference laboratory.
Manufacturer
Low Bilirubin Level
High Bilirubin Level
SKML Labs
SKML Labs
February
September
March 2010
February
September
March 2010
Roche
101
99
96
106
103
101
Beckman
112
113
115
108
108
109
Olympus
113
108
Abbott
105
106
106
110
106
106
Ortho Vitros
77
77
94
86
78
87
Siemens Dimension
101
97
98
102
104
105
Siemens Advia
109
92
108
111
110
109
Green: percentages from 96 – 104% Amber: percentages from 91 -95 and 105 – 109% Red: percentages <91 and >109%
•
di-taurobilirubine en BSA (gebruikt in de meeste commerciële kalibratoren) interfereren in de bilirubinebepaling
•
UBIL en TBIL worden toegevoegd aan verschillende eiwitmatrices
•
meting op 7 analysers met diazo methoden
•
in bovine sera worden veel lagere concentraties gevonden dan in humane sera
•
verschil sterker in commerciële bovine preparaten dan in abattoir serum. Inhibitoren diazoreactie / acceleratoren?
•
Onvoldoende recovery DTB in humaan serum door Vitros en Architect
•
In verse bovine matrix (abattoir) 5 van de 7 analysers goed
•
In de commerciële bovine sera laten alle analysers een te lage recovery van DTB zien
•
Auteurs gebruiken reeds 20 jaar BSA, nooit problemen
•
Inhibitoren uit voeding of bij processing gebruikte containers? Azide? Aging? auteurs vinden geen verklaring
•
“The solution to this problem is simple: Use human instead of bovine serum for preparing bilirubin calibrators”
Resultaten SKML neonatale bilirubine 2010.1
SKML 2010.B
GEM
SD
VC
µmol/l
µmol/l
%
2010.1A
215
15
7.2
2010.1B
414
26
6.3
2010.1C
524
38
7.2
2010.1D
405
25
6.3
2010.1E
214
13
6.2
2010.1F
312
18
5.8
2010.1D
n
GEM
SD
VC
µmol/l
µmol/l
%
Diazo-p-nitroaniline
8
409
12
2.8
Jendrassik-Grof. Autom.
23
407
25
6.0
DiCl-phenyl diaz. Zout
25
403
18
4.4
Jendrassik-Grof achtigen
63
405
20
4.8
Ictometer
8
419
14
3.2
spectrofotometrisch
12
418
13
3.1
(DPD)
Met dank aan dr. Cas Weykamp Ing. Riejean Kuylaars dr. Joke Apperloo
Low µmol/L
30
Combi 2009-1
25
Target Reference Method
20
60
70
Roche
OCD
Beckman
Siemens-Dade Behring
Abbott
Siemens-Bayer
80
High µmol/L
Low µmol/L
25
Combi 2010-1
20
Target Reference Method
15
65
75
Roche
OCD
Beckman
Siemens-Dade Behring
Abbott
Siemens-Bayer
85
High µmol/L