bs_bs_banner

Clinical and Experimental Immunology O R I G I N A L A RT I C L E doi:10.1111/cei.12369

Calculated globulin (CG) as a screening test for antibody deficiency

S. Jolles,* R. Borrell,* S. Zouwail,†‡ Summary

A. Heaps,* H. Sharp,* M. Moody,*
Calculated globulin (total protein – albumin) is usually tested as part of a
C. Selwood,* P. Williams,*
liver function test profile in both primary and secondary care and deter-
C. Phillips,§ K. Hood,¶ S. Holding**

Downloaded from https://academic.oup.com/cei/article/177/3/671/6421257 by guest on 07 August 2022

mines the serum globulin concentration, of which immunoglobulins are a
and T. El Shanawany*
Departments of *Immunology and †Biochemistry,
major component. The main use hitherto of calculated globulin is to detect
University Hospital of Wales, ¶School of Medicine, paraproteins when the level is high. This study investigated the potential to
Cardiff University, Cardiff, §Swansea Centre for use low levels of calculated globulin to detect antibody deficiency. Serum
Health Economics, Swansea University, Swansea, samples with calculated globulin cut-off < 18 g/l based on results of a pilot
**Immunology Department, Hull Royal study were collected from nine hospitals in Wales over a 12-month period.
Infirmary, Hull, UK, and ‡Medical Biochemistry Anonymized request information was obtained and the samples tested for
Department, School of Medicine, Alexandria immunoglobulin levels, serum electrophoresis and, if appropriate,
University, Alexandria, Egypt immunofixation. A method comparison for albumin measurement using
bromocresol green and bromocresol purple was undertaken. Eighty-nine per
cent (737 of 826) samples had an immunoglobulin (Ig)G level of < 6 g/l
using the bromocresol green methodology with a cut-off of < 18 g/l, and 56%
(459) had an IgG of < 4 g/l. Patients with both secondary and primary anti-
body deficiency were discovered and serum electrophoresis and
immunofixation showed that 1·2% (10) had previously undetected small
paraproteins associated with immune-paresis. Using bromocresol purple,
74% of samples had an IgG of < 6 g/l using a cut-off of < 23 g/l. Screening
using calculated globulin with defined cut-off values detects both primary
and secondary antibody deficiency and new paraproteins associated with
immune-paresis. It is cheap, widely available and under-utilized. Antibody-
deficient patients have been discovered using information from calculated
Accepted for publication 27 April 2014 globulin values, shortening diagnostic delay and time to treatment with
Correspondence: S. Jolles, Department of immunoglobulin replacement therapy.
Immunology, University Hospital of Wales,
Cardiff CF14 4XW, Wales, UK. Keywords: common variable immunodeficiency, immunoglobulin, myeloma,
E-mail: [email protected] primary antibody deficiency, screening, secondary antibody deficiency

the frequent requirement for medical attention and inter-

Introduction
vention. The morbidity and need for health resources is
Antibody deficiency may be either primary or secondary, exacerbated by a diagnostic delay of, on average, between 6
and the most common severe primary antibody deficiency and 7 years before appropriate treatment is started [2]. In
(PAD) is common variable immunodeficiency (CVID), contrast to many other PID diseases, patients with CVID
which accounts for the majority (57%) of all symptomatic may be diagnosed at almost any age [2], which has implica-
primary immunodeficiency (PID) cases on the European tions for screening approaches.
Society for Immunodeficiencies (ESID) registry (http:// The misconception that PID always presents in infancy in
www.esid.org). Secondary antibody deficiency occurs even combination with the protean manifestations of CVID, the
more frequently than PAD. number of specialities to which patients may present and
CVID is an important diagnosis not only because of its the wide range of non-infectious complications all contrib-
prevalence (one in 25 000–50 000) [1], but also because of ute to the delay in recognition of the underlying diagnosis.

© 2014 The Authors. Clinical & Experimental Immunology published by John Wiley & Sons Ltd on behalf of British Society 671
for Immunology, Clinical and Experimental Immunology, 177: 671–678
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution
and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
S. Jolles et al.

This may result in recurrent infections leading to irrevers- indicate haematological malignancy, such as multiple
ible end organ damage, such as bronchiectasis. myeloma. The globulin fraction consists of alpha 1, alpha 2,
Clinical strategies to reduce diagnostic delay include beta 1, beta 2 and gamma globulin fractions, and as proteins
improved education and awareness of antibody deficiency within the alpha and beta fractions can act as acute-phase
in both primary and secondary care, patient-centred screen- proteins; high levels of these could potentially mask a low
ing [3], the use of the ‘10 warning signs’ of PID (http:// gamma globulin (antibody) level. CG levels have been
www.info4pi.org) and using computer sorting of diagnostic shown to be low in patients with antibody deficiency [11],
codes [4]. A study into the clinical features which identify and as they reflect antibody levels there is an opportunity to
children with immunodeficiency found that of the 10 aid clinicians in the diagnosis of a range of conditions with
warning signs the strongest predictors were family history, both high and low levels (Table 1).
intravenous antibiotics for sepsis (neutrophil defects) and In preparation for this study, a survey of practice in bio-
failure to thrive (T cell defects), while for B cell defects the chemistry laboratories across Wales and a pilot study into
only predictor was family history of immune deficiency [5]. the interpretation of calculated globulin were undertaken.

Downloaded from https://academic.oup.com/cei/article/177/3/671/6421257 by guest on 07 August 2022

Laboratory approaches have involved opportunistic LFTs are performed in very high numbers (1·9 million/year
detection cases of antibody deficiency by cascade testing for for a population of 3·06 million in Wales) with total protein
immunoglobulins when there is a very low background in already included. Even if not tested as standard within the
enzyme-linked immunosorbent assays (ELISAs) used for LFT profile the inclusion of total protein is inexpensive,
coeliac disease testing, such as the immunoglobulin (Ig)A costing £0·07, US$0·11 or €0·08 per test.
anti-tissue transglutaminase (TTG) assay [6], and also CG also fulfils the accepted criteria for a screening test
when there is a very low IgE < 2 IU/ml [7]. These opportun- based on those developed originally by the World Health
istic methods lack specificity and depend upon the clinical Organization (WHO), which encompass the condition, the
details on the request form, which are often incomplete or test and the treatment [12]. These state that the condition
lacking. Furthermore, the very successful newborn screen- should be an important health problem, the natural history
ing for severe combined immunodeficiency (SCID) looking should be understood and there should be a detectable risk
for T cell receptor excision circles (TRECs) in DNA extracted factor, disease marker, recognizable latent or early sympto-
from Guthrie spots [8], when extended to B cell kappa- matic stage. The test should be simple, safe, precise and vali-
deleting recombination excision circles (KRECs) [9], will not dated with a defined cut-off level. The test should be
detect CVID due to the adult onset and presence of B cells acceptable to the population and there should be an agreed
in the majority of patients. Guthrie spots are also unsuitable policy on the further investigation of individuals with a
for the detection of IgA deficiency in neonates, as the IgA at positive result. There should be an effective treatment or
that age is of both fetal and maternal origin [10]. There still intervention for patients identified through early detection,
remains a major challenge in the early detection of B cell with evidence of early, rather than late, treatment leading to
immunodeficiency and thus antibody deficiency. better outcomes.
Calculated globulin (CG) is derived from the difference In the setting of PID, such as CVID, immunoglobulin
between total protein and albumin results, and forms part replacement remains the mainstay of treatment, reduces
of the liver function test (LFT) profile. CG is used widely in infection frequency and severity and increases life expec-
primary and secondary care to detect high levels that may tancy. Higher doses of immunoglobulin are associated with

Table 1. Causes of a high or low calculated globulin.

High calculated globulin levels Low calculated globulin levels
Plasma cell dyscrasias and lymphoproliferative disease, e.g. myeloma, Primary antibody deficiency, e.g. CVID
Waldenstrom’s macroglobulinaemia
Monoclonal gammopathy of undetermined significance (MGUS) Small paraproteins with immune-paresis
Viral infections, e.g. HIV, EBV, HBV Secondary antibody deficiency, e.g. lymphoma, chemotherapy, rituximab,
protein losing enteropathy, plasma exchange, renal losses (e.g. nephrotic
syndrome) and immunosuppressive drugs
Sarcoidosis Fluid therapy/shifts, e.g. peri-operatively or in acutely unwell patients
such as ITU or medical therapy involving fluid resuscitation
Connective tissue disease, e.g. systemic lupus erythematosus, Sjögrens
syndrome, rheumatoid arthritis
Autoimmune hepatitis, primary biliary sclerosis, primary sclerosing
cholangitis
Chronic infection, e.g. TB, osteomyelitis, endocarditis, hepatitis
EBV = Epstein–Barr virus; HBC = hepatitis B core antigen; CVID = common variable immunodeficiency; TB = tuberculosis; ITU = intensive
care unit.

672 © 2014 The Authors. Clinical & Experimental Immunology published by John Wiley & Sons Ltd on behalf of British Society
for Immunology, Clinical and Experimental Immunology, 177: 671–678
 CG as a screen for antibody deficiency

reduced infection frequency [13,14]. Late diagnosis and Following a subsequent change in the measurement
delayed institution of immunoglobulin replacement method of albumin from BCG (Abbott Architect) to
therapy results in increased morbidity and mortality bromocresol purple (BCP) (Abbott Architect) in our centre,
[15,16]. There is now a large range of immunoglobulin a further study was performed to identify the method-
treatment options for patients comprising intravenous specific CG cut-off using 25 anonymized samples for each
(IVIg), subcutaneous (SCIg) weekly or biweekly, facilitated CG level from 19 g/l to 26 g/l, as there was predicted to be a
subcutaneous (fSCIg) and rapid push [13,17–19] made pos- change in the albumin value between the methods. Results
sible by improvements in immunoglobulin manufacturing, were analysed using the ‘Analyse it Method Validation’
resulting in more concentrated products, faster infusion edition in Microsoft Excel and receiver operating character-
rates using IVIg and for fSCIg the development of recombi- istic (ROC) curves used to identify the most appropriate
nant human hyaluronidase (rhuPH20) [19,20]. cut-off for the BCP method.

Methods Results

Downloaded from https://academic.oup.com/cei/article/177/3/671/6421257 by guest on 07 August 2022

A survey was carried out via the Wales Biochemistry audit The survey of practice with respect to use and interpreta-
group using a questionnaire to determine current practice tion of CG results was performed with responses from all
with respect to the use of calculated globulin (CG) across 13 biochemistry laboratories in Wales (population 3·06
the 13 biochemistry laboratories in Wales. All question- million), and the findings showed that approximately 1·9
naires were returned and the data analysed using Microsoft million LFTs were performed across Wales, with 0·82
Excel. million of these in primary care. Although analytical plat-
An analysis of anonymized adult (> 16 years) samples forms were different between the laboratories, all were
with a range of CG levels from 15 to 22 g/l was undertaken using the BCG method for measuring serum albumin and
for 50 samples at each level of CG (Abbott Diagnostics, the Biuret method for total protein measurement. Con-
Maidenhead, UK) using the colorimetric bromocresol green versely, there was no consistent reference range for CG,
(BCG) method for albumin and the Architect Biuret varying at the upper end of the range from 35 to 45 g/l
method (16200 Abbott Analyzer; Abbott Diagnostics) for and at the lower end from ‘no lower limit’ to 20–28 g/l. At
total protein (TP). Immunoglobulin levels were then meas- the lower end of the range, seven of the 10 laboratories
ured by nephelometry (Siemens BN2 Nephelometer; reporting CG either took no action or did not report low
Siemens, Camberley, UK) to confirm the relationship levels of CG, while two flagged low results and one
between CG and IgG and to identify an appropriate cut-off reviewed results if < 25 g/l (three laboratories did not
value for subsequent study. Testing was carried out in clini- report CG).
cal pathology-accreditated (CPA) laboratories. The pilot study showed a linear relationship between
Based on the results of the pilot study, ethical and values of IgG and CG with a wide range of IgG levels for
research and development committee approval was each level of CG (Fig. 1), with a CG of 18 g/l corresponding
obtained for a more detailed study of samples having CG with an IgG of 6 g/l, the lower limit of the normal range
levels of < 18 g/l from nine participating hospitals in Wales (6–16 g/l). The cut-off value of CG < 18 g/l was chosen
[University Hospital of Wales (UHW), University Hospital which corresponded to a sensitivity of 0·82 and a specificity
Llandough, Velindre Hospital, Royal Gwent Hospital, of 0·71 for an IgG < 3 g/l (Table 2). Following the presenta-
Neville Hall Hospital, Caerphilly Miners Hospital, Bronglais tion of these data to the biochemistry department a number
Hospital, West Wales General Hospital and Prince Charles of samples with low CG were referred to immunology for
Hospital]. Samples collected from adults (age > 16 years) immunoglobulin levels, and at least one patient was diag-
during a period of 1 year were anonymized and sent to nosed with PID and commenced on replacement immuno-
UHW for assay of immunoglobulin levels (IgG, IgA and globulin therapy. However, audit 1 year later of 100 random
IgM) (Siemens BN2 Nephelometer; Siemens), serum elec- samples with a CG of <18 g/l showed no action/comments
trophoresis (Sebia Capillarys 2; Sebia, Norcross, GA, USA) on any sample from biochemistry.
and, where appropriate, serum immunofixation (Sebia In the study of serum samples with CG < 18 g/l from
Hydrasys; Sebia). The laboratory and clinical data were ana- nine hospitals in Wales, after exclusion of duplicates (on the
lysed using Microsoft Excel and Graphpad Prism version 6. same patient), 826 samples were tested for immunoglobulin
Duplicate samples were identified through the patient ini- levels and serum electrophoresis with the addition of
tials, sex, age, source or location of request and clinical immunofixation to confirm and type any paraproteins. The
details, and removed from further analysis. The initial pres- results demonstrated that 89% of samples had an IgG of
ence of duplicate samples, however, resulted in improved < 6 g/l (adult normal range 6–16 g/l) and 56% had an
capture of clinical details. The results were presented to the IgG < 4 g/l with the distribution shown in Fig. 2, support-
biochemistry department at UHW and reporting practice ing the chosen cut-off value of < 18 g/l using the BCG
audited 1 year later. method. The BCP method, using a cut-off of < 23 g/l for

© 2014 The Authors. Clinical & Experimental Immunology published by John Wiley & Sons Ltd on behalf of British Society 673
for Immunology, Clinical and Experimental Immunology, 177: 671–678
S. Jolles et al.

15 200
180
160

Number of samples
140
120
100
10 80
60
40
IgG (g/l)

20
0

Downloaded from https://academic.oup.com/cei/article/177/3/671/6421257 by guest on 07 August 2022

Ig <3 l
Ig <4 l
Ig <5 l
Ig <6 l
Ig <7 l
Ig <8 l
Ig <9 /l
Ig <10 /l
Ig <11 /l
Ig <12 /l
Ig <13 /l
Ig <14 /l
Ig <15 /l
<1 g/l
l
G g/
G g/
G g/
G g/
G g/
G g/

g/
G g
g
g
g
g
g
g
Ig <2

6
G
5

G
G
G
G
G
G
G
Ig
Level of IgG

Fig. 2. Numbers of samples at each level of immunoglobulin (Ig)G.

The numbers of samples at each level of IgG are shown for samples
with a calculated globulin (CG) of < 18 g/l with only 11% of samples
recorded with an IgG level in the normal range (6–16 g/l); 56% of
samples had an IgG of < 4 g/l, which is 4 standard deviations below
0 the mean for adults [bromocresol green (BCG) method].
14 16 18 20 22 24
Calculated globulin (g/l)

Fig. 1. Correlation of calculated globulin levels with immunoglobulin CG, had a sensitivity of 0·76 and specificity of 0·54 at an
(Ig)G. Fifty samples obtained anonymously for each level of calculated IgG level of < 3 g/l (Table 2).
globulin (CG) at 15, 16, 17, 18, 19, 20, 21 and 22 g/l were tested for The sources of samples with an IgG < 6 g/l are shown in
immunoglobulin levels and the correlation with IgG is shown. An IgG Fig. 3, with a wide range of clinical specialities and primary
of 6 g/l corresponded to a CG level of 18 g/l and this was selected as care represented. It is noteworthy that when the stringency
the cut-off value for the bromocresol green (BCG) method.
in terms of significant antibody deficiency is increased by
using an IgG level of < 4 g/l there is a major fall in samples
from surgery of 122 to 20, perhaps suggesting that some of

Table 2. Comparison of sensitivity and specificity using BCG and BCP methods.
IgG < 3 g/l IgG < 4 g/l IgG < 5 g/l
Sensitivity Specificity Sensitivity Specificity Sensitivity Specificity
Calculated globulin cut-off for BCG (g/l)
< 16 0·24 0·95 0·15 0·95 0·11 0·96
< 17 0·53 0·83 0·45 0·84 0·39 0·87
< 18 0·82 0·71 0·78 0·74 0·66 0·78
< 19 0·94 0·58 0·90 0·61 0·81 0·65
< 20 1·00 0·41 0·98 0·44 0·92 0·48
< 21 1·00 0·28 0·98 0·29 0·95 0·32
< 22 1·00 0·14 1·00 0·14 0·99 0·16
Calculated globulin cut-off for BCP (g/l)
< 20 0·24 0·89 0·26 0·92 0·21 0·92
< 21 0·48 0·78 0·51 0·83 0·45 0·87
< 22 0·64 0·66 0·66 0·71 0·63 0·77
< 23 0·76 0·54 0·79 0·59 0·77 0·65
< 24 0·84 0·41 0·87 0·45 0·85 0·50
< 25 0·96 0·28 0·94 0·31 0·92 0·35
< 26 1·00 0·14 1·00 0·16 1·00 0·20
Sensitivity and specificity of variable cut-off for immunoglobulin (Ig) G below set value when using bromocresol green (BCG) and bromocresol
purple (BCP) methods for albumin and Biuret method for total protein.

674 © 2014 The Authors. Clinical & Experimental Immunology published by John Wiley & Sons Ltd on behalf of British Society
for Immunology, Clinical and Experimental Immunology, 177: 671–678
 CG as a screen for antibody deficiency

180 antibody deficiency in patients being monitored for immu-

nosuppressive therapy and with a known diagnosis of CLL,
160
but also in patients with diabetes, which is more difficult to
140 explain, and with a range of non-specific symptoms such as
Number of samples

120 tiredness and bloating.

100

80

60

40 Leukaemia Waldenstroms
MDS
20 CLL

Downloaded from https://academic.oup.com/cei/article/177/3/671/6421257 by guest on 07 August 2022

0
Not given
Medicine
Haematology
Oncology
GP
OPD
ITU
Cardiac
Renal
Gastro
Surgery
Neurology
TCTU
Immunology
Paeds
Cardiac ITU
A&E
Rheumatology
MAU
Trauma
Maternity
Lymphoma ALL

CML

Source of sample Myeloma

Fig. 3. Source of samples with an immunoglobulin (Ig)G < 6 g/l. A

wide range of specialities is represented with a high number from
surgery due probably to the use of intravenous fluids and blood Amyloid
products which may temporarily alter a calculated globulin (CG)
Fig. 4. Haematology diagnoses with IgG < 4 g/l. The largest category
[intensive care unit (ITU), general practice (GP), outpatients
of antibody-deficient patients from haematology is from requests
department (OPD), Teenage Cancer Trust unit (TCTU), accident and
labelled lymphoma [lymphoma, non-Hodgkin lymphoma, diffuse
emergency (A&E), medical assessment unit (MAU].
large B cell lymphoma (DLBCL), follicular lymphoma, Hodgkin
lymphoma, renal lymphoma, Mantle cell lymphoma, marginal zone
lymphoma (MZL), Burkitt’s lymphoma and central nervous system
(CNS) lymphoma] and a range of other haematological diagnoses
the changes associated with intravenous replacement of
reflecting either secondary antibody deficiency due to the disorder or
fluid or blood result in more limited reductions in antibody subsequent to its treatment.
levels.
The speciality with the highest number of samples with
low IgG apart from surgery was haematology, and the associ-
ated disorders with an IgG < 4 g/l are shown in Fig. 4. Unex-
pectedly, the largest diagnostic group was lymphoma, at 34%,
rather than CLL at 20%; however, it is possible that some of Sulphasalazine
3% Chemotherapy
Gold
the CLL patients may already be receiving immunoglobulin, Statin 3% 3%
Collapse 3% Methotrexate
which may lower the numbers of CLL patients with antibody IHD 3% 10%
deficiency detected by CG. The results identify substantial 3% Clozapine
CCF 13%
numbers of undiagnosed secondary antibody-deficient 3%
patients, including some with neutropenic sepsis or pneumo- CLL/Lymphoma
9%
nia stated in the clinical information on the request form. A N/A
19%
similar pattern is seen with unrecognized antibody deficiency
in samples from oncology. Abdo Tired
bloating 6% Diabetes
In addition to the antibody deficiency, in 1·2% of samples 3% 16%
(10 of 826) a new, previously undetected small paraprotein Cough
3%
with immune-paresis was identified, with IgG levels ranging
from 1·93–4·8 g/l. The associated immune-paresis suggests
Fig. 5. Diagnoses with immunoglobulin (Ig)G < 4 g/l from primary
that these patients may no longer fall into the category of
care. Diagnoses from 32 request forms from patients in general
monoclonal gammopathy of undetermined significance practice with IgG levels less than 4 g/l are shown. There are a
(MGUS) and would merit referral to haematology for significant number where drug monitoring is taking place and a
further assessment and potential intervention. number with a range of more non-specific symptoms. Congestive
Clinical details on samples originating from primary care cardiac failure (CCF), ischaemic heart disease (IHD), chronic
with IgG levels < 4 g/l are shown in Fig. 5. These highlight lymphocytic leukaemia (CLL) and information not available (NA).

© 2014 The Authors. Clinical & Experimental Immunology published by John Wiley & Sons Ltd on behalf of British Society 675
for Immunology, Clinical and Experimental Immunology, 177: 671–678
S. Jolles et al.

The impact of reduction in diagnostic delay on wellbeing, nasal drip and joint pains had settled. Prior to replacement
schooling, end-organ damage and the use of resources is immunoglobulin he rated his wellbeing as three out of 10 on
demonstrated by the patient histories of three cases identified a visual analogue scale, and in the year following commence-
initially by low CG levels who were seen following the study. ment of immunoglobulin this improved to eight out of 10,
and he missed no days from work.
Case 1
Case 3
A 19-year-old male was referred to immunology on the basis
of low CG. His mother recalled him constantly dribbling in A 26-year-old female working for the university had
infancy and she was told ‘his tongue was too big for his a 9-year history of recurrent infections, including
mouth’. He had tonsillectomy and adenoidectomy aged 4 pneumonias requiring chest X-rays, intravenous antibiotics
years and three insertions of grommets and a hearing aid. He and admissions to hospital on three occasions. She felt she
had a chronic cough from the age of 4 years and subsequently ‘has a constant chest infection for 6 months of the year’ and

Downloaded from https://academic.oup.com/cei/article/177/3/671/6421257 by guest on 07 August 2022

sinusitis and post-nasal drip. He suffered from 12–15 infec- had been diagnosed as having irritable bowel syndrome. In
tions annually and lost 50–60 days of school per year. He was the previous year she had lost 21 days from work, needed
sent home from school as his coughing was disrupting the six courses of antibiotics and rated her wellbeing as 6 out of
class, and recalled falling asleep in a geography examination 10 on a visual analogue scale. In the previous 9 years she
as he had been coughing all night. When asked when he last had had 72 visits to her GP and had been told she needed to
felt well he said ‘never’, and rated his wellbeing as 3 out of explore psychological causes for her symptoms. Her CG was
10. In the past year he had six courses of antibiotics, two 18 g/l, with pan-hypogammaglobulinaemia (IgG < 1·34 g/l,
admissions with pneumonia and had investigations for HIV IgA < 0·24 g/l, IgM < 0·17 g/l) and a poor response to vacci-
and cystic fibrosis. During the previous 15 years it was esti- nation. She was diagnosed with CVID and commenced on
mated he would have visited his general practitioner (GP) on Subgam 6·75 g per week subcutaneously, and had a single
180 occasions, and in 2012 he was found to have pan- visit to the GP in the year following the start of treatment,
hypogammaglobulinaemia with absent vaccine responses had no infections, required no antibiotics and felt the treat-
and was diagnosed as having CVID. Since commencing ment had had a ‘brilliant effect on her quality of life’.
replacement immunoglobulin 18 months ago he has required
no further admissions to hospital.
Discussion
There are no published studies of the utility of CG to screen
Case 2
for antibody deficiency, and the principal findings of the
A 23-year-old mechanic was referred to gastroenterology current study are that using appropriately established
with blood in his stools aged 21 years and was investigated method-dependent CG cut-off values it is possible to detect
with endoscopies and diagnosed as having a distal colitis of primary and secondary antibody deficiencies, as well as
indeterminate type. At this time his CG was 14, and treat- small new paraproteins associated with immune-paresis.
ment for colitis was commenced (Asacol) with little benefit. This approach fulfils the criteria of a screening test [12],
He was investigated for coeliac disease when a duodenal and as a proof-of-principle, patients detected with low CG
biopsy showed normal villous architecture with a lack of have been diagnosed with CVID and commenced on
plasma cells, his CG was 15 and results showed IgA defi- immunoglobulin replacement. Using the BCG methodology
ciency. In 2011, on the basis of the low CG, immunoglobu- for albumin measurement and the Biuret method for total
lin levels were tested (IgG < 1·34 g/l, IgA < 0·06 g/l, protein [21], a cut-off value of < 18 g/l for calculated globu-
IgM < 0·17 g/l) and he was referred to immunology. At this lin defines a population in which 89% have an IgG level of
time he had had watery or semi-formed motions eight < 6 g/l.
times per day for 3–4 years, suffered from joint pains, was A limitation of this study is that it did not evaluate the
coughing an eggcup of light green sputum per day for the utility of CG in children, as the normal ranges for immuno-
past year, with sinusitis and a post-nasal drip, and had lost, globulins vary with age; however, levels approximating the
on average, 40 days of work per year. He was found to have adult ranges are achieved by the age of 9–10 years [22] and
low numbers of class-switched memory B cells, did not potentially further studies could address whether screening
respond to test vaccination with pneumovax II and so was using CG could be used from 10 years of age. Clinically,
diagnosed with CVID. children below this age have a higher frequency of infection
Immunoglobulin replacement therapy with Privigen 50 g which declines with age.
intravenously every 3 weeks was commenced, and at his most There are several approaches to the introduction of elec-
recent follow-up his colitis had resolved with two formed tronic comments to aid clinicians. The most straightfor-
motions per day and his previous colitis medication was dis- ward would be on samples from primary care, as the use of
continued. He no longer had a cough and his sinusitis, post- intravenous fluids (which may temporarily perturb CG

676 © 2014 The Authors. Clinical & Experimental Immunology published by John Wiley & Sons Ltd on behalf of British Society
for Immunology, Clinical and Experimental Immunology, 177: 671–678
 CG as a screen for antibody deficiency

levels) in this setting would be unlikely, and the patients in potential utility of CG. In chronic lymphocytic leukaemia
the case histories all reported frequent contact with their (CLL), 30–50% of deaths are due to infection caused by sec-
GP. Several comments were developed for a low CG in ondary antibody deficiency, T cell defects and the effects of
primary care and the GP laboratory management commit- therapy [25].
tee liaison group selected the following comment on the Variations in detection rate may be expected, as the
basis that it interprets the underlying abnormality and sug- populations being tested will be different for different labo-
gests action which may be considered in the appropriate ratories and health settings. The presence of a large haema-
clinical setting: ‘Low calculated globulin may represent anti- tology department and oncology centre will have influenced
body deficiency. Consider immunoglobulin measurement if the results in terms of secondary antibody deficiency in this
there is a history of infections’. study. The ability of CG to be used as a screening test
In secondary care a similar comment could be used, with should be widely applicable, regardless of the presence of
an added caveat that intravenous fluid administration in major hospitals or specialist services. Given the low costs of
surgical, medical or intensive care may cause a temporarily testing, use of CG may also be relevant in resource-poor

Downloaded from https://academic.oup.com/cei/article/177/3/671/6421257 by guest on 07 August 2022

low CG. settings.
Some laboratories could also choose to cascade directly The information presented by this study could be used to
on discovery of a low CG, adding immunoglobulins and inform health policy regarding the most appropriate test
SEP as deemed appropriate; however, this may be hampered profiles to use. There is the potential to reduce diagnostic
by the frequent lack of clinical data on the request form and delay, optimize the use of health resources and reduce the
the requesting clinician may be in the best position to make burden of disease not only for patients with antibody defi-
the most informed decision, potentially also limiting costs. ciency but also a much wider range of conditions. CG is an
With increasing pressure on health resources, the removal underutilized test which is cheap (£0·07, US$0·11 or €0·08),
of total protein has been advocated on cost grounds. As part widely available and fulfils the need for a screening
of the process of test profile harmonization and demand approach for antibody deficiency.
management [23] it has been suggested that ‘its inclusion is
more as an opportunistic screening test for monoclonal
gammopathy, but no specific intervention exists other than Acknowledgements
watchful waiting’ [24]. S. J. and A. H. are supported by NISCHR Fellowships.
Data from the current study, however, argue that the Support for consumables was provided in a grant from CSL
range of disorders for which additional information is Behring. The support of the Wales Biochemistry Audit
gained is wider than monoclonal gammopathy (Table 1); in group is acknowledged in gathering information about
particular, when defined cut-offs are in place, there are clear current practice for CG in Wales.
treatment advances for multiple myeloma, clinically impor-
tant monoclonal bands are detected with a low CG and
screening criteria are fulfilled for both primary and second- Disclosures
ary antibody deficiency. In addition, testing costs must be
evaluated in the context of the complete patient pathway, T. E. has received payment from Biotest and CSL Behring
the utility losses associated with delay in the diagnosis of for attendance at advisory board meetings. S. J. has received
patients, the benefits associated with accurate immunologi- support from CSL Behring, Baxter, Biotest, BPL and
cal testing and the potentially increased use of health Octapharma for projects, advisory boards, meetings and
resources if the test is not performed. Where the test is clinical trials.
already being performed it would seem reasonable to obtain
best value for money by education regarding its use at both
References
ends of the reference range.
There may be circumstances when an antibody deficiency 1 Jolles S. The variable in common variable immunodeficiency: a
does not require immunoglobulin replacement, but further disease of complex phenotypes. J Allergy Clin Immunol Pract
investigation, monitoring and treatment may still be altered 2013; 1:545–56.
in the light of the result and the clinician and patient are 2 Cunningham-Rundles C. How I treat common variable immune
deficiency. Blood 2010; 116:7–15.
able to proceed in a more informed manner.
3 de Vries E. Patient-centred screening for primary immunodefi-
The increasing use of B cell ablation therapy across a
ciency, a multi-stage diagnostic protocol designed for non-
range of specialities (in particular haematology) and the
immunologists: 2011 update. Clin Exp Immunol 2012; 167:
growth of maintenance therapy with rituximab, alongside 108–19.
the development of a range of agents targeting B cells for 4 Cunningham-Rundles C, Sidi P, Estrella L, Doucette J. Identifying
malignant, autoimmune and inflammatory indications, is undiagnosed primary immunodeficiency diseases in minority
likely to result in an increase in secondary antibody defi- subjects by using computer sorting of diagnosis codes. J Allergy
ciency in a subset of these patients, further increasing the Clin Immunol 2004; 113:747–55.

© 2014 The Authors. Clinical & Experimental Immunology published by John Wiley & Sons Ltd on behalf of British Society 677
for Immunology, Clinical and Experimental Immunology, 177: 671–678
S. Jolles et al.

5 Subbarayan A, Colarusso G, Hughes SM et al. Clinical features 15 Wood P, Stanworth S, Burton J et al. Recognition, clinical diagno-
that identify children with primary immunodeficiency diseases. sis and management of patients with primary antibody deficien-
Pediatrics 2011; 127:810–6. cies: a systematic review. Clin Exp Immunol 2007; 149:410–23.
6 Bright P, Lock RJ, Unsworth DJ. Immunoglobulin A deficiency on 16 Aghamohammadi A, Moin M, Farhoudi A et al. Efficacy of intra-
serological coeliac screening: an opportunity for early diagnosis of venous immunoglobulin on the prevention of pneumonia in
hypogammaglobulinaemia. Ann Clin Biochem 2012; 49 (Pt patients with agammaglobulinemia. FEMS Immunol Med
5):503–4. Microbiol 2004; 40:113–8.
7 Unsworth DJ, Virgo PF, Lock RJ. Immunoglobulin E deficiency: a 17 Berger M, Murphy E, Riley P, Bergman GE. Improved quality of
forgotten clue pointing to possible immunodeficiency? Ann Clin life, immunoglobulin G levels, and infection rates in patients with
Biochem 2011; 48 (Pt 5):459–61. primary immunodeficiency diseases during self-treatment with
8 Verbsky JW, Baker MW, Grossman WJ et al. Newborn screening subcutaneous immunoglobulin G. South Med J 2010; 103:856–63.
for severe combined immunodeficiency; the Wisconsin experience 18 Shapiro RS. Subcutaneous immunoglobulin: rapid push vs.
(2008–2011). J Clin Immunol 2012; 32:82–8. infusion pump in pediatrics. Pediatr Allergy Immunol 2013;
9 Borte S, von Dobeln U, Fasth A et al. Neonatal screening for severe 24:49–53.

Downloaded from https://academic.oup.com/cei/article/177/3/671/6421257 by guest on 07 August 2022

primary immunodeficiency diseases using high-throughput 19 Wasserman RL, Melamed I, Stein MR et al. Recombinant human
triplex real-time PCR. Blood 2012; 119:2552–5. hyaluronidase-facilitated subcutaneous infusion of human immu-
10 Borte S, Janzi M, Pan-Hammarstrom Q et al. Placental transfer of noglobulins for primary immunodeficiency. J Allergy Clin
maternally-derived IgA precludes the use of guthrie card eluates as Immunol 2012; 130:951–7 e11.
a screening tool for primary immunodeficiency diseases. PLOS 20 Jolles S. Hyaluronidase facilitated subcutaneous immunoglobulin
ONE 2012; 7:e43419. in primary immunodeficiency. ImmunoTargets Ther 2013; 2:125–
11 Arnold DF, Wiggins J, Cunningham-Rundles C, Misbah SA, 33.
Chapel HM. Granulomatous disease: distinguishing primary anti- 21 Corcoran RM, Durnan SM. Albumin determination by a modified
body disease from sarcoidosis. Clin Immunol 2008; 128:18–22. bromocresol green method. Clin Chem 1977; 23:765–6.
12 Wilson JMG, Jungner G. Principles and practice of screening for 22 Milford Ward A, Sheldon J, Rowbottom A, Wild GD, eds. Hand-
disease. Public Health Papers 34: 1–163. Geneva: World Health book of clinical immunochemistry, 9th edn. Sheffield: PRU Publi-
Organization (WHO), 1968. cations, 2007.
13 Orange JS, Belohradsky BH, Berger M et al. Evaluation of correla- 23 Fryer AA, Smellie WS. Managing demand for laboratory tests: a
tion between dose and clinical outcomes in subcutaneous immu- laboratory toolkit. J Clin Pathol 2013; 66:62–72.
noglobulin replacement therapy. Clin Exp Immunol 2012; 24 Smellie WS. Time to harmonise common laboratory test profiles.
169:172–81. BMJ 2012; 344:e1169.
14 Orange JS, Grossman WJ, Navickis RJ, Wilkes MM. Impact of 25 Francis S, Karanth M, Pratt G et al. The effect of immunoglobulin
trough IgG on pneumonia incidence in primary immunodefi- VH gene mutation status and other prognostic factors on the inci-
ciency: a meta-analysis of clinical studies. Clin Immunol 2010; dence of major infections in patients with chronic lymphocytic
137:21–30. leukemia. Cancer 2006; 107:1023–33.

678 © 2014 The Authors. Clinical & Experimental Immunology published by John Wiley & Sons Ltd on behalf of British Society
for Immunology, Clinical and Experimental Immunology, 177: 671–678