Technical Briefs
Prolactin Results for Samples Containing Macroprolactin Are Method and Sample Dependent, Georges Gilson,1*
Patrick Schmit,1 Jean Thix,2 Jean-Paul Hoffman,3 and RenéLouis Humbel1 (1 Laboratoire de Biochimie et
d’Immunopathologie, Centre Hospitalier de Luxembourg, rue Barblé 4, 1210 Luxembourg, Luxembourg;
2
Laboratoire Clinique Sainte-Marie, rue Wurth-Paquet 7,
4350 Esch-sur-Alzette, Luxembourg; 3 Laboratoire National de Santé, Division de Biochimie, rue du Laboratoire
42, 1911 Luxembourg, Luxembourg; * author for correspondence: fax 352-457794, e-mail gilson.georges@chl.lu)
Prolactin (PRL) exists in human serum in several molecular forms that can be identified by gel-filtration chromatography (GFC). The 23-kDa monomer is the predominant form in the general population, but other circulating
species include the 50-kDa form (big PRL) and the 150- to
170-kDa macroprolactin (big big PRL) (1, 2 ). The prevalence of macroprolactin in the general population is
currently unknown, but this macromolecular form of PRL
has been characterized as a complex of PRL with an IgG
antibody (3 ) that has reduced bioactivity in vivo (4 ) and
a variable reactivity with commercial immunoassays for
PRL (4 – 8 ). Macroprolactin is cleared from the blood
circulation more slowly than monomeric PRL, leading to
an apparent hyperprolactinemia, depending on the immunoassay used for the measurement of PRL. Thus,
identification of macroprolactin as a cause of high PRL in
a patient sample is important because it can help resolve
diagnostic confusion and avoid expensive pituitary imaging studies and inappropriate treatment (5 ). As a result of
distribution of serum containing macroprolactin in the
United Kingdom National External Quality Assessment
Scheme, immunoassays for the measurement of PRL have
been subdivided into three classes according to their
reactivity with macroprolactin: low-, medium-, and highreading methods (6 ).
We used the polyethylene glycol (PEG) precipitation
method (4 ) to screen for the presence of macroprolactinemia in a population of 319 samples with increased serum
PRL (.30 mg/L) measured by our routine method [electrochemiluminescence immunoassay (Elecsys 2010;
Roche)], an immunoassay that reacts strongly with macroprolactin (high-reading method). The presence of macroprolactin was confirmed by GFC on a Sephacryl S-300
column (Pharmacia) as described elsewhere (4 ). All samples containing macroprolactin, as well as 48 hyperprolactinemic samples containing exclusively monomeric
PRL, were additionally assayed for PRL by two other
commercially available automated immunoassay systems:
a medium-reading method for macroprolactin [chemiluminescent immunoassay (Immulite; Diagnostic Products
Corporation)] and a low-reading method [chemiluminescent immunoassay (ACS:180; Bayer)].
PEG precipitation has been validated as a screening
method for the presence of macroprolactin only when
used with the Wallac Delfia PRL assay (4 ). In our experiments with PEG precipitation (250 g/L PEG 6000 at room
temperature, freshly prepared PEG reagent every 3
months) and the Elecsys PRL assay, we obtained the
following results: over a 2-week period, 10 repeated
determinations of PRL recovery after PEG precipitation of
a sample containing exclusively monomeric PRL and a
sample containing 82% macroprolactin gave mean recoveries of 84% (SD, 5%) and 14% (SD, 0.8%), respectively. To
validate a cutoff for PEG precipitation with the Elecsys
PRL assay, we used a subset of our hyperprolactinemic
samples for an initial study. In 30 hyperprolactinemic
samples with no evidence of macroprolactin by GFC, the
recovery of PRL after PEG precipitation was 53–96%,
whereas in 30 samples in which macroprolactin had been
identified recovery was 6 – 46%. Consequently, when PRL
was measured with the Elecsys assay, recovery of ,50%
of PRL after PEG precipitation was considered a positive
screening result for macroprolactin. The PEG precipitation method, however, could not be used with the two
other automated PRL immunoassays because the presence of PEG produced negative interference with the
ACS:180 assay (4 ) and positive interference with the
Immulite assay (recovery .100% after PEG precipitation
for samples containing monomeric PRL).
In our population of 319 hyperprolactinemic samples,
59 sera (18%) gave a PRL recovery of ,50% after PEG
precipitation, and in all of these samples, the presence of
macroprolactin was confirmed by GFC. This finding is
consistent with the data of other authors reporting a
prevalence of 15.4 –26% for macroprolactinemia among
hyperprolactinemic samples (4, 7 ). Fig. 1A compares the
Elecsys and the ACS:180 PRL assays. The results obtained
with the Elecsys assay were higher than those measured
with the ACS:180 assay, and the difference was influenced
by the presence of macroprolactin. The results for sera
containing only monomeric PRL were, on average, 40%
higher with the Elecsys assay (range, 18 – 60%) than with
the ACS:180 assay, whereas the results for samples containing macroprolactin were on average 78% higher
(range, 20 –143%). The samples for which the difference
between the Elecsys and the ACS:180 results was .60%
were exclusively macroprolactinemic samples. Of the 59
samples, 17 (29%) containing macroprolactin had a difference between the Elecsys and the ACS:180 results of
20 –59%, which was similar to the difference observed in
the monomeric PRL population.
The difference plot for the Elecsys and the Immulite
assays (Fig. 1B) shows that the Elecsys assay gave higher
results than the Immulite assay and that the bias was
influenced by the presence of macroprolactin. In the
Elecsys assay, the sera containing only monomeric PRL
gave results that were, on average, 44% higher (range,
14 –79%) than in the Immulite assay, and the macroprolactinemic samples gave results that were 78% higher
(range, 46 –130%) on average. All samples presenting a
difference .80% between the Elecsys and the Immulite
results were macroprolactinemic samples, and all samples
presenting a difference ,45% were samples containing
only monomeric PRL. The overlapping of the macroprolactinemic and the monomeric PRL populations was
greater than for the Elecsys/ACS:180 comparison because
Clinical Chemistry 47, No. 2, 2001
331
332
Technical Briefs
56% (33 of 59) of the macroprolactinemic samples present
a difference of the Elecsys and the Immulite results
between 45% and 80%.
Fig. 1C compares the medium-reading Immulite
method and the low-reading ACS:180 method. Although
in some samples the results by the two methods were
quite different, there was little overall bias attributable to
the presence of macroprolactin. We obtained an average
difference of 25.1% (range, 251% to 30%) for the monomeric PRL samples and an average difference of 2.4%
(range, 250% to 86%) for the macroprolactin-containing
samples.
We confirmed in our study that, in general practice,
macroprolactinemia is a common phenomenon in hyperprolactinemic samples. Of the three automated immunoassay systems tested, the Elecsys assay gave higher results
than those obtained by the Immulite and the ACS:180
assays, and the bias was influenced considerably by the
presence of macroprolactin. Our data confirmed that the
reactivity with macroprolactin is dependent on the assay
used for the PRL measurement, but we additionally
showed that the PRL results obtained for samples containing macroprolactin are also sample dependent. Thus,
even if the average difference between results of macroprolactinemic samples measured with the Elecsys and
ACS:180 assays (78%) was much higher than the average
difference for monomeric PRL samples (40%), we could
identify a subpopulation of sera containing macroprolactin that behave like monomeric samples and present only
a difference of 20 –59%. Consequently, it is not possible to
exclude the presence of macroprolactin by comparison of
the results obtained by a high-reading method (Elecsys
2010; Roche) and a low-reading method (ACS:180; Bayer),
as has been proposed by some authors (9 ). In such a
comparison, 29% of the macroprolactinemic samples of
our population would not have been recognized. The
difference plot illustrating the Immulite/ACS:180 comparison (Fig. 1C) shows little overall bias attributable to
the presence of macroprolactin, but the considerable
range in the differences between the Immulite and the
Fig. 1. Difference plots of PRL results using samples containing
macroprolactin (F) or monomeric PRL (E).
(A), Elecsys/ACS:180 comparison. (B), Elecsys/Immulite comparison. (C), Immulite/ACS:180 comparison.
Clinical Chemistry 47, No. 2, 2001
ACS:180 results obtained for macroprolactinemic samples
(range, 250% to 86%) indicates a highly variable, sampledependent response of the ACS:180 assay to macroprolactin. The great disparity of values observed when comparing results from macroprolactinemic samples measured
by the Elecsys or the Immulite assay with the results
obtained by a low-reading method such as ACS:180 may
reflect variation in the structure of macroprolactin. Macroprolactin is most probably not one unique macromolecule but rather a heterogeneous family of PRL-IgG complexes that react differently depending on the type of
immunoassay used for PRL determination.
In conclusion, our study reinforces the point that PRL
assays from different manufacturers give highly variable
prolactin results for samples containing macroprolactin
(4 – 8 ). Our data additionally show that the reactivity of
macroprolactin in a PRL immunoassay, be it a low-,
medium-, or high-reading method, is not identical for all
macroprolactinemic samples. This finding underscores
the necessity of a systematic screening strategy for macroprolactin in all samples with increased PRL
(4, 5, 10, 11 ). With the Elecsys PRL assay, PEG precipitation, with a cutoff value of 50%, was an efficient and
easy-to-use screening tool for the presence of macroprolactin. Because of the interference of PEG in some commercially available PRL assays, the confirmation of macroprolactinemia may require time-consuming methods,
such as centrifugal ultrafiltration (9 ) and GFC.
We thank Michael N. Fahie-Wilson, Department of
Clinical Chemistry, Southend Hospital, Westcliff-on-Sea,
Essex, UK, for expert advice.
References
1. Suh HK, Frantz AG. Size heterogeneity of human prolactin in plasma and
pituitary extracts. J Clin Endocrinol Metab 1974;39:928 –35.
2. Smith CR, Norman CR. Prolactin and growth hormone: molecular heterogeneity and measurement in serum. Ann Clin Biochem 1990;27:542–50.
3. Cavaco B, Leite V, Santos MA, Arranhado E, Sobrinho LG. Some forms of big
big prolactin behave as a complex of monomeric prolactin with an immunoglobulin G in patients with macroprolactinemia or prolactinoma. J Clin
Endocrinol Metab 1995;80:2342– 6.
4. Fahie-Wilson MN, Soule SG. Macroprolactinaemia: contribution to hyperprolactinaemia in a district general hospital and evaluation of a screening test
based on precipitation with polyethylene glycol. Ann Clin Biochem 1997;34:
252– 8.
5. John R, McDowell IFW, Scanlon MF, Ellis AR. Macroprolactin reactivities in
prolactin assays: an issue for clinical laboratories and equipment manufacturers [Letter]. Clin Chem 2000;46:884 –5.
6. Fahie-Wilson MN, Ellis AR, Seth J. Macroprolactin—a major problem in
immunoassays for prolactin [Abstract]. Clin Chem 1999;45(Suppl 6):A83.
7. Bjoro T, Morkrid L, Wergeland R, Turter A, Kvistborg A, Sand T, Torjesen P.
Frequency of hyperprolactinaemia due to large molecular weight prolactin
(150 –170 kD PRL). Scand J Clin Lab Invest 1995;55:139 – 47.
8. Cavaco B, Prazeres S, Santos MA, Sobrinho LG, Leite V. Hyperprolactinemia
due to big big prolactin is differently detected by commercially available
immunoassays. J Endocrinol Invest 1999;22:203– 8.
9. Fahie-Wilson MN, Ellis AR. Macroprolactin—what should we do about it?
Proc UKNEQAS Meeting 1998;3:121–3.
10. Lindstedt G. Endogenous antibodies against prolactin—a “new” cause of
hyperprolactinemia. Eur J Endocrinol 1994;130:439 – 42.
11. Vieira JGH, Tachibana TT, Obara LH, Maciel RMB. Extensive experience and
validation of polyethylene glycol precipitation as a screening method for
macroprolactinemia [Technical Brief]. Clin Chem 1998;44:1758 –9.
333
Multiplexed Mutagenically Separated PCR: Simultaneous Single-Tube Detection of the Factor V R506Q
(G1691A), the Prothrombin G20210A, and the Methylenetetrahydrofolate Reductase A223V (C677T) Variants,
Georg Endler,1 Paul A. Kyrle,2 Sabine Eichinger,2 Markus
Exner,1 and Christine Mannhalter1* (1 Department of Laboratory Medicine, Molecular Biology Division, and 2 University Clinic for Internal Medicine 1, Department of
Hematology, AKH Wien, Währinger Gürtel 18-20, 1190
Wien, Austria; * author for correspondence: e-mail
christine.mannhalter@univie.ac.at)
Recently, mutations in several genes that encode for
coagulation proteins, such as the factor V (FV):R506Q
(G1691A) mutation or the prothrombin (FII):G20210A
variant, have been identified as important risk factors for
developing a venous thromboembolism (VTE) (1 ). These
mutations contribute to the development of thrombosis in
;50% of all patients.
The FV:R506Q (G1691A) mutation currently is considered the most important genetic risk factor for venous
thrombosis. Although 5% of the healthy population carry
the mutant allele, the prevalence of this mutation in
patients with venous thrombosis is ;40%. Compared
with the wild type, heterozygous carriers of the mutation
have an 8-fold higher risk and homozygotes have an 80to 100-fold higher risk of developing a VTE (2 ). If and
how this mutation contributes to arterial thrombosis is
still under investigation, although recently Ardissino et al.
(3 ) showed correlations among the factor V mutation,
smoking, and myocardial infarction. The FII:G20210A
variant, which is associated with increased prothrombin
activity and increased plasma prothrombin, has been
identified as an independent risk factor for thrombosis
(4 ). Compared with individuals with the wild-type genotype, heterozygous carriers of the mutation have a 2.8- to
5-fold higher risk of developing VTE. The influence of the
FII:G20210A mutation on arterial thromboembolic disease
is not clear, but there is evidence that it is associated with
cerebrovascular ischemic disease (5 ) and myocardial infarction (6 ).
In addition to these two established thrombotic risk
factors, the role of other genetic variations is still under
investigation. The A223V (C677T) mutation in the thermolabile methylenetetrahydrofolate reductase (MTHFR)
gene is associated with mild hyperhomocysteinemia. Homozygosity for the mutation may be associated with an
increased risk for cardiovascular events. The effect of the
mutation on venous thrombosis is still controversial (7–
10 ).
For diagnostic analyses and for scientific studies of
large numbers of patients, fast and economic assays that
can be performed with standard PCR instruments are
highly desirable.
We developed a mutagenically separated (MS) multiplex PCR for the simultaneous detection of mutations in
the FV, FII, and MTHFR genes (11 ). Our MS-PCR is a
single-tube PCR-based technique using allele-specific
primers that differ in length by 8 –10 bp. Base mismatches