Thermal degradation of polyethylene glycol 6000 and its
effect on the assay of macroprolactin
Louise Boughena, 1, John Liggatb and *Graham Ellisa,
a
Department of Clinical Biochemistry, St. John's Hospital, Livingston, West Lothian, EH54 6PP, Scotland
b
Department of Pure and Applied Chemistry, Strathclyde University, Glasgow G1 1XL, Scotland
* Corresponding author. Clinical Biochemistry, St. John's Hospital, Livingston, West Lothian, EH54 6PP, Scotland.
Fax: +44 1506 523355
Clinical Biochemistry
Volume 43, Issue 9, June 2010, Pages 750-753
doi:10.1016/j.clinbiochem.2010.02.012
Received 24 November 2009;
revised 26 February 2010;
accepted 27 February 2010.
Available online 19 March 2010.
Originally presented as an abstract to Focus 2009, the National Meeting of The Association for Clinical Biochemistry,
Liverpool, 19th–21st May 2009; Annals of Clinical Biochemistry 2009;46(Suppl. 1):47(abst.).
Abstract
Objectives
To study the effectiveness of partially degraded polyethylene glycol 6000 (PEG) as a precipitant for macroprolactin.
Design and methods
PEG was heated to 63 °C in air for up to 20 days and its effectiveness assessed as a precipitant for sera containing
normal prolactin or macroprolactin. Decomposition was studied chemically and with NMR spectroscopy.
Results
Thermal degradation was similar to what had occurred over several years of natural degradation. Initially PEG
degraded 2–5 days caused excess precipitation of monomeric prolactin (false-positive macroprolactinemia).
Samples degraded 18–20 days failed to precipitate macroprolactin, giving false negative results. Two 1H NMR peaks
at 4–4.5 ppm were not detectable in undegraded PEG but were after 1 day. Their relative integral increased to
20 days.
Conclusions
Aging of PEG can be accelerated by heating. The suitability of PEG for use in macroprolactin assays can be
assessed by the absence of peaks at 4–4.5 ppm by 1H NMR.
Keywords: Prolactin; Polyethylene glycols; Humans; Reproducibility of results; Magnetic resonance spectroscopy;
Temperature
Introduction
Serum macroprolactin is a high molecular mass form of prolactin that clears less quickly from serum, increasing
serum total prolactin [1]. Because prolactin is measured as a marker of pituitary function, especially to detect
prolactinomas, elevated values caused by the presence of macroprolactin can lead to diagnostic confusion.
Macroprolactin is quite common and was found in 3.7% of a group of Japanese hospital workers; in most cases the
macroprolactin was a complex of prolactin and prolactin autoantibodies [2].
Polyethylene glycol 6000 (PEG) is used in Clinical Biochemistry as a precipitant for macroprolactin. It is generally
regarded as stable in crystalline form but it can be degraded with heat and steam to facilitate its decomposition [3].
Ellis described a problem with an old reagent lot of PEG that had degraded in crystalline form in the reagent
container [4]. Initially, degradation caused the reagent to precipitate monomeric prolactin as well as macroprolactin in
serum samples, leading to false-positive macroprolactinemia. Later in the degradation process, PEG samples failed
to precipitate serum macroprolactin, leading to false-negative macroprolactinemia.
Polymer chemists study the accelerated decomposition of relatively stable polymers by heating above room
temperature in air [3] and [5]. Han et al. [6] studied PEG decomposition at 80 °C in air over 42 days. Decomposition
products at high temperature may vary from those associated with room temperature decomposition. In the present
study, decomposition at 63 °C was used as this was the minimum temperature for practical accelerated
decomposition. The effectiveness of PEG as a precipitant for macroprolactin was studied over the course of this
decomposition. The process was monitored colorimetrically and with proton NMR spectrometry. The aim was to
assess criteria for PEG quality that could assure its suitability for clinical assays of macroprolactin.
Methods
Patient samples
Left-over samples from sera that had been tested for prolactin and macroprolactin were used. The project was
approved by the Lothian Research Ethics Committee. Surplus UKNEQAS quality control samples were also used.
These had been tested by gel filtration and were kindly provided by Mr. Andy Ellis of UKNEQAS, Edinburgh.
Degradation of PEG and its effect on macroprolactin assays
PEG was obtained from VWR International BDH Prolabo, product number 442714 K. The reagent Batch Number
was 06 L140003. The reagent was used within 18 months of the Date Packed and 2–4 months before the date until
which the specificity of the unopened product was guaranteed. Initial experiments showed that crystalline PEG was
relatively stable below its melting point (53–63 °C, depending on source)—data not shown. So PEG was heated at
63 °C for up to 20 days. Duplicate samples of 1 or 4 g of crystalline PEG were used in 10 or 60 mL clear glass
powder bottles with the black plastic-lined screw cap removed to allow air exposure. One of each degraded duplicate
samples was reconstituted with water to give 250 g/L PEG and they were used to assay serum containing
monomeric prolactin or macroprolactin on an Abbott ARCHITECT Analyzer [4]. The second sample was used for 1H
NMR and reaction with Schiff's reagent.
Procedure for assay of macroprolactin
Briefly 0.2 mL of serum was treated with either 0.2 mL Abbott ARCHITECT Multi-Assay Manual Diluent (Product
Reference 7D82-50, phosphate buffered saline solution) or 250 g/L PEG solution. The mixed solutions were left to
stand for 10 min; then centrifuged at 10,500g in an Abbott Microcentrifuge. The supernatants were transferred to
Abbott ARCHITECT sample cups and assayed for prolactin. After correcting for dilution, the results were expressed
as percentage recovery by dividing the result in the PEG supernatant by that of the control. Monomeric samples
recover 60–90%; equivocal ones 40–60%; and those with significant macroprolactin < 40% [7].
NMR spectroscopy
PEG (0.1 g) was dissolved in 1.5 mL deuterated chloroform (Sigma-Aldrich product #151823) and
13C
and 1H NMR
spectra prepared on a Bruker ADVANCE/DPX400 400 MHz instrument to monitor degradation (oxidation) of the
samples.
Colorimetric assay of aldehydes
Aldehydes were tested colorimetrically by mixing 150 µL of 250 g/L PEG solutions with 2.0 mL deionised water and
1.0 mL of Schiff's reagent (VWR International BDH Prolabo, product number 30969.261). Samples were left at room
temperature for 30 min then the absorbance was read at 560 nm. Results were compared against a calibration curve
prepared by diluting Molecular Biology Grade formaldehyde solution (372 g/L), (Calbiochem cat #344198) to give 0–
1.0 mmol/L. The detection limit was 0.002 mmol/L. This was determined as 2 Standard Deviations of 15 replicate
measurements of a calibrator with a zero concentration of formaldehyde.
Results
Visual appearance of the PEG crystals and the protein precipitation
At 63 °C, the PEG melted. When examined after cooling, crystalline structure was visible within the melt in samples
degraded 1–2 days but by 2–20 days, the crystals had coalesced and looked like slightly turbid ice. The later
samples were slower to dissolve in water, having a slightly oily or greasy appearance before eventually dissolving.
They appeared to give less protein precipitate than untreated samples when used to precipitate macroprolactin from
serum.
Degradation of PEG and its effect on macroprolactin assays
Initially PEG solutions degraded 2–5 days at 63 °C caused excess precipitation of monomeric prolactin, giving falsepositive macroprolactinemia. In contrast, samples degraded for 18–20 days failed to precipitate macroprolactin,
giving false negative results for macroprolactin-containing sera (Fig. 1).
Fig. 1. Recovery of prolactin and macroprolactin with PEG solutions degraded for different times. Initially the effect of
degradation is greatest on monomeric prolactin with under-recovery causing misclassification as “macroprolactin.”
After PEG has degraded longer, it is ineffective in precipitating macroprolactin or monomeric prolactin.
NMR spectroscopy
13C
NMR spectra were uninformative, being similar in PEG and degraded PEG. A large peak at 3.6 ppm was visible
on all 1H spectra, corresponding to the 1H signal of the polyethylene backbone CH2 groups. Over the degradation,
other small peaks (or groups of peaks) developed at 2.5, 4.0–4.5, 5.4, 7.3, 8.1 and 9.7 ppm. Two groups of peaks at
4–4.5 ppm were absent in native PEG but they appeared in association with the degradation (Fig. 2). The relative
integral (peak area/area of biggest peak) progressively increased. At 1 day they represented a relative integral of
0.00148. This increased to 0.0241 at 20 days over the course of the degradation (Fig. 3).
Fig. 2. PEG 1H NMR spectra on different days. Note the appearance of peaks at 4.0–4.5 ppm.
Fig. 3. Relative integral of 4.0–4.5 ppm peaks with degradation time.
Colorimetric assay of aldehydes
Schiff-reacting products increased over the course of the PEG degradation (Fig. 4). The colour increase is probably
due to formaldehyde or other aldehydes released over the course of the degradation.
Fig. 4. Increasing Schiff-reacting products with degradation time.
Discussion
Physical appearance of the PEG
Heating the PEG caused it to melt and it seemed to lose its crystalline appearance in later samples. PEG that had
degraded naturally in the reagent bottle initially appeared indistinguishable from fresh PEG (white crystals) even
when it was ineffective in macroprolactin assays [4]. After a few more years in the capped bottle, the degraded PEG
had a strong aldehyde/ketone odour and it had become a paste (like toothpaste). Proton NMR changes at 4–4.5 ppm
were similar but more marked than those we observed in PEG denatured at 63 °C.
Degradation of PEG and its effect on macroprolactin assays
With respect to the effect of thermal degradation of PEG, this seemed to mirror what had happened over several
years of natural degradation. Initially the degraded samples caused false positive macroprolactinemia, whereas later
they failed to precipitate macroprolactin, leading to false negative macroprolactinemia.
Products of the degradation process
From the 1H NMR spectra, decomposition products appear to contain the structures consistent with a free-radical
oxidation. Han et al. [6] heated PEG to 80 °C and observed similar peaks that they attributed low molecular weight
acid esters such as those of formic acid following breakdown of the PEG to substances with shorter chain lengths.
Unlike Han et al., we did not observe useful information in the
13C
spectra. Decomposition of polyethoxylated
alcohols (used in cosmetics) underwent a similar free radical decomposition to produce formaldehyde and peroxides
[8]. Reaction with Schiff's reagent suggested some aldehyde products were produced at 63 °C.
Nature of the degradation process
Degradation at 63 °C may not be identical to natural degradation in a closed reagent container at room temperature.
Volatile degradation products, such as formaldehyde or formic acid esters could be lost at 63 °C but retained at room
temperature. They may contribute to excess protein precipitation in macroprolactin assays. We used open containers
for degradation so as to not restrict the oxygen available for degradation but this could lead to loss of volatiles. PEG
(4 g) is 0.7 mmol; a 60 mL closed bottle of air contains approximately (60/22400) × (1/5) = 0.5 mmol oxygen. (This
approximation is made from the Molar Volume of an ideal gas under standard conditions and the fraction of oxygen
in air). So there would be approximately only about one molecule of oxygen per molecule of polymer (Molecular
Mass = 6000) in a closed container.
Restricted oxidation of PEG has been observed a vacuum [6] and in closed containers (John Liggat and Hugh
Berkmeyer, personal communication). These latter authors found that there were only a limited number of volatiles
captured at −196 °C when PEG samples that had been heated to 63 °C for 6 weeks in sealed jars were subjected to
sub-ambient temperature distillation. Mass spectrometry identified these as water, carbon dioxide, 2-hydroxyethyl
formate and an unidentified aromatic compound.
Nonetheless NMR spectroscopy does seem to correlate with changes in PEG due to degradation and with its
effectiveness in macroprolactin assays.
Laboratories should be aware that PEG may decompose in the reagent container over several years. Current
reagents have an Expiration Date that should not be exceeded. On the basis of these accelerated decomposition
experiments, the suitability of PEG for use as a laboratory reagent for macroprolactin assays can be assessed by the
absence of peaks at 4–4.5 ppm by 1H NMR. We recognize that NMR instrumentation is not standard clinical
laboratory equipment. However, one of us (G.E) has advocated that kit manufacturers should supply suitably
certified PEG solutions as part of their diagnostic kits (4). It seems irrational that diagnostic companies produce
prolactin kits with care and great attention to regulatory guidelines, yet their kits may contribute to serious
misdiagnosis of prolactinomas when macroprolactin is present [2] and [9].
Acknowledgments
We thank Mr. Andy Ellis of UKNEQAS, Edinburgh Royal Infirmary, for providing serum with monomeric and
macroprolactin.
References
[1] M.N. Fahie-Wilson, R. John and A.R. Ellis, Macroprolactin; high molecular mass forms of circulating prolactin,
Ann Clin Biochem 42 (Pt 3) (May 2005), pp. 175–192.
[2] N. Hattori, T. Ishihara and Y. Saiki, Macroprolactinaemia: prevalence and aetiologies in a large group of hospital
workers, Clin Endocrinol (Oxf) (May 28 2009).
[3] E. Otal, D. Mantzavinos, M.V. Delgado, R. Hellenbrand, J. Lebrato and I.S. Metcalfe et al., Integrated wet air
oxidation and biological treatment of polyethylene glycol-containing wastewaters, J Chem Technol Biotechnol 70 (2)
(1997), pp. 147–156
[4] G. Ellis, Degradation of crystalline polyethylene glycol 6000 and its effect on assays for macroprolactin and other
analytes, Clin Biochem 39 (10) (Oct 2006), pp. 1035–1040.
[5] R.J. Young and P.A. Lovell, Introduction to polymers (2 ed.), Chapman and Hall, New York (1991).
[6] S. Han, C. Kim and D. Kwon, Thermal/oxidative degradation and stabilization of polyethylene glycol, Polymer 38
(2) (1997), pp. 317–323
[7] L. Germano, M. Migliardi, D. Marranca, A. Mormile, L. Filtri and E. Cocciardi et al., Evaluation of polyethylene
glycol precipitation as screening test for macroprolactinemia using Architect immunoanalyzer (abstract), Clin Chim
Acta 355 (Suppl 5) (2005), p. S262.
[8] M. Bergh, K. Magnusson, J.L. Nilsson and A.T. Karlberg, Formation of formaldehyde and peroxides by air
oxidation of high purity polyoxyethylene surfactants, Contact Dermat 39 (1) (Jul 1998), pp. 14–20.
[9] A.M. Suliman, T.P. Smith, J. Gibney and T.J. McKenna, Frequent misdiagnosis and mismanagement of
hyperprolactinemic patients before the introduction of macroprolactin screening: application of a new strict laboratory
definition of macroprolactinemia, Clin Chem 49 (9) (2003), pp. 1504–1509.