Download Use of the lymphocyte count as a diagnostic screen in adults with

The Laryngoscope
C 2013 The American Laryngological,
V
Rhinological and Otological Society, Inc.
Use of the Lymphocyte Count as a Diagnostic Screen in Adults With
Suspected Epstein–Barr Virus Infectious Mononucleosis
Timothy C. Biggs, MBBCh, MRCS, DOHNS; Stephen M. Hayes, MBBS, BSc, MRCS, DOHNS;
Jonathan H. Bird, BMBS, MRCS, DOHNS; Philip G. Harries, MBBS, FRCS (ORL-HNS);
Rami J. Salib, BMedSci, BMBS, DLO, PhD, FRCS (ORL-HNS)
Objectives/Hypothesis: To evaluate the predictive diagnostic accuracy of the lymphocyte count in Epstein–Barr virus–
related infectious mononucleosis (IM).
Study Design: Retrospective case note and blood results review within a university-affiliated teaching hospital.
Methods: A retrospective review of 726 patients undergoing full blood count and Monospot testing was undertaken.
Monospot testing outcomes were compared with the lymphocyte count, examining for significant statistical correlations.
Results: With a lymphocyte count of 4 3 109/L, 99% of patients had an associated negative Monospot result (sensitivity of 84% and specificity of 94%). A group subanalysis of the population older than 18 years with a lymphocyte count 4
3 109/L revealed that 100% were Monospot negative (sensitivity of 100% and specificity of 97%). A lymphocyte count of
4 3 109/L correlated significantly with a negative Monospot result.
Conclusions: A lymphocyte count of 4 3 109/L appears to be a highly reliable predictor of a negative Monospot
result, particularly in the population aged >18 years. Pediatric patients, and adults with strongly suggestive symptoms and
signs of IM, should still undergo Monospot testing. However, in adults with more subtle symptoms and signs, representing
the vast majority, Monospot testing should be restricted to those with a lymphocyte count >4 3 109=L.
Key Words: Infectious mononucleosis, glandular fever, tonsillitis, Epstein–Barr virus, lymphocyte, diagnosis.
Level of Evidence: NA
Laryngoscope, 123:2401–2404, 2013
INTRODUCTION
Glandular fever was first described in 1889 as a
condition comprised of fever, pharyngitis, cervical
lymphadenopathy, and lymphocytosis.1 The term infectious mononucleosis (IM) was coined in 1920 after
Sprunt and Evans described patients with this condition
as having lymphocytosis with atypical lymphocytes.1
The exact cause of IM remained unknown until the
1960s, when the Epstein–Barr virus (EBV) was discovered; this is now accountable for the majority of cases of
IM worldwide.2 Rarer causes of IM include cytomegalovirus, human herpes virus 6, toxoplasmosis and human
immunodeficiency virus.2 In 1932, Paul and Bunnell
devised the first heterophile antibody test, after it was
From the University Hospital Southampton National Health Service Foundation Trust, Southampton (T.C.B., S.M.H., P.G.H., R.J.S.), United
Kingdom, Portsmouth Hospital National Health Service Trust, Portsmouth (J.H.B.), United Kingdom
Editor’s Note: This Manuscript was accepted for publication
January 14, 2013.
Presented at the South West ENT Academic Meeting, Bath,
United Kingdom, June 8, 2012 (winner of the best oral presentation
prize) and Otorhinolaryngological Research Society Autumn Meeting,
Norwich, United Kingdom, September 7, 2012.
The authors have no other funding, financial relationships, or conflicts of interest to disclose.
Send correspondence to Mr. Timothy Biggs, ENT Department,
University Hospital Southampton NHS Foundation Trust, Tremona
Road,
Southampton,
SO16
6YD,
United
Kingdom.
E-mail:
[email protected]
DOI: 10.1002/lary.24030
Laryngoscope 123: October 2013
observed that serum from patients with IM caused sheep
erythrocytes to agglutinate.2 This then became the most
widely used test for the detection of IM and was further
validated by Hoagland in 1975 (Hoagland criteria).3
According to Hoagland criteria, a diagnosis of IM can be
made in the presence of fever, pharyngitis, and lymphadenopathy with at least a 50% lymphocytosis, of which
10% should be atypical together with confirmation of a
positive serological test.3 The Monospot test has now
replaced the Paul–Bunnell test, most commonly using
latex particles together with an antigen to cause agglutination, thereby confirming the presence of heterophile
antibodies produced in EBV infection.2 Current commercially available Monospot tests are quoted to be 70% to
92% sensitive and 96% to 100% specific.4 At the University Hospital Southampton National Health Service
Foundation Trust (UHS), where this study was undertaken, the Mono-Latex test (Biokit, Barcelona, Spain),
reported to be 93.9% sensitive and 98.3% specific for the
presence of EBV, is currently in use.4–6
Evidence within the literature suggests that the
lymphocyte to white cell count ratio could be used as an
adjunct when assessing the probability of EBV-related
IM while awaiting further serological testing.5,6 In many
cases, this formal serological testing is costly and can
take up to 48 hours to provide a result, depending on
the laboratory. We present a much easier, more specific,
and cheaper alternative in assessing the probability of
EBV-related IM in those presenting with signs and
Biggs et al: Lymphocyte Count as Diagnostic Screen
2401
TABLE I.
Statistical Correlation of Total and Differential White Cell Counts With Monospot Results.
Monospot Positive
Monospot Negative
Mean
SD
CI
Mean
SD
CI
Mann–Whitney U
Test Significance
7.1
1.4
3.5
0.6
6.1-8.1
1.3-1.6
2.3
0.7
1.5
0.4
2.1-2.4
0.7-0.7
<.0001
<.0001
Neutrophil count, 109/L
2.9
1.5
2.5-3.3
4.9
3.0
4.6-5.1
<.0001
Total white cell count, 109/L
11.9
4.2
10.7-13.1
8.0
3.5
7.8-8.3
<.0001
Count
Lymphocyte count, 109/L
Monocyte count, 109/L
CI 5 confidence interval; SD 5 standard deviation.
symptoms of IM, thus reducing the need for routine
Monospot testing in a large cohort of patients.
performed in SPSS version 20 (SPSS, Chicago, IL) using an independent sample Mann–Whitney U test (significance achieved
at P <.05).
MATERIALS AND METHODS
RESULTS
A total of 726 patients with Monospot and FBC
results were included within the study. Of the 726
patients studied, 50 (7%) were Monospot positive. The
mean age in the Monospot-positive group was 21 years
(range 5 12–64 years), with a mean age of 30 years
(range 5 8 months–93 years) in the Monospot-negative
group. Table I highlights the mean values, standard
deviation, and confidence intervals of total and differential white cell counts compared to Monospot results,
together with the calculated statistical correlations.
The differences between the Monospot-positive and
Monospot-negative groups were all highly statistically
significant (P <.0001).
Table II outlines the positive and negative predictive values, specificity, and sensitivity of the total and
differential white cell counts in relation to positive and
negative Monospot results. Of all the indices measured,
the lymphocyte count alone possessed the highest negative predictive value (99%) and highest specificity (94%).
All indices showed relatively low positive predictive values (1%–53%), and associated sensitivities ranged from
2% (neutrophil count) to 84% (lymphocyte count). Based
on these observations, and from the pathophysiological
standpoint of the disease process, the lymphocyte count
appeared to represent the most appropriate index for
the purposes of this study.
Study Design
The laboratory results for 749 retrospective patients presenting with sore throat and lymphadenopathy, undergoing
Monospot testing at UHS between October 1, 2011 and January
20, 2012, were included within the study. Clinical symptoms of
sore throat, fever, and lymphadenopathy were present for all
samples studied, ascertained by assessing the requesting information on corresponding blood forms. Twenty-three cases without full blood count (FBC) results were excluded from the study.
The patient population included general practice and UHS
(inpatients, outpatients, and emergency department), as well as
a small number of patients from outlying hospitals.
Monospot results (positive or negative) were compared
with their associated lymphocyte, monocyte, neutrophil, and
total white cell counts, which were classified as positive (above
the normal reference range for the UHS laboratory) or negative
(within the normal reference range for the UHS laboratory).
The cutoff values for the normal reference ranges within the
laboratory at UHS were as follows: lymphocyte count 4 3 109/
L, monocyte count 1 3 109/L, neutrophil count 7.5 3 109/L,
and total white cell count 11 3 109/L. Monospot testing outcomes were compared with FBC indices including lymphocyte,
monocyte, neutrophil, and total white cell counts, examining for
any significant statistical correlations.
Statistical Analysis
The data (nonparametric) was collated and analyzed using
Excel 2009 (Microsoft, Redmond, WA) with statistical analysis
TABLE II.
Predictive Values of the Total and Differential White Cell Counts in Relation to Monospot Results.
Monospot Positive
Count
Above
Reference
Range
Within
Reference
Range
Monospot Negative
Above
Reference
Range
Within
Reference
Range
Positive
Predictive
Value
Negative
Predictive
Value
Sensitivity
Specificity
Lymphocyte count, reference
range 4 3 109/L
42
8
38
638
53%
99%
84%
94%
Monocyte count, reference
range 1 3 109/L
36
14
74
602
33%
98%
72%
89%
Neutrophil count, reference
range 7.5 3 109/L
1
49
83
593
1%
92%
2%
88%
Total white cell count, reference
range 11 3 109/L
26
24
87
589
23%
96%
52%
87%
Laryngoscope 123: October 2013
2402
Biggs et al: Lymphocyte Count as Diagnostic Screen
Comparison With Other Studies
TABLE III.
Absolute Lymphocyte Count Predictive Values.
Monospot Result
Totals
Predictive
Value
38
638
80
646
53% PPV
99% NPV
50
676
726
84%
sensitivity
94%
specificity
Lymphocyte
Count, 109/L
>4
4
Totals
Sensitivity/
specificity
Positive
Negative
42
8
NPV 5 negative predictive value; PPV 5 positive predictive value.
Table III outlines the positive and negative predictive value, specificity, and sensitivity of the lymphocyte
count in relation to a positive and negative Monospot
result. Table IV highlights the results of a group subanalysis of the predictive value of the lymphocyte count
in relation to the Monospot results in the population
aged >18 years. The overall negative predictive value of
the lymphocyte count is 99% in the group as a whole.
However, when examining patients older than 18 years,
this value and the sensitivity figure both rose to 100%.
DISCUSSION
Summary
This study highlights the lymphocyte count as a
very sensitive (84%) and specific (94%) marker, with a
high predictive value (99%) of a negative Monospot test.
These values are even more impressive when examining
patients older than 18 years, with an associated sensitivity of 100%, sensitivity of 97%, and negative predictive
value of 100%.
The clinical implications and application of these
results are significant. In adult patients presenting with
fever, sore throat, and cervical lymphadenopathy, a FBC
should be obtained before undertaking routine Monospot
or serological testing. If the lymphocyte count is 4 3
109=L, there is a 99% probability that the patient will
have a negative Monospot test. If the patient is >18
years of age, then this figure rises to 100%. Overall, the
study findings support the limitation of routine Monospot testing in suspected cases of EBV-related IM to
adult patients with a lymphocyte count >4 3 109=L,
pediatric patients, or where there exists a strong clinical
suspicion of IM.
Within the literature, there have been a number of
studies evaluating the use of the lymphocyte count in
relation to the total white cell count in predicting the
occurrence of EBV-related IM.5,6 These studies compared
the ratio of the lymphocyte count to the total white cell
count, quoting a value of 0.35 as the most specific and
sensitive marker for the presence of EBV-related IM. This
is reported to provide a sensitivity of between 84% and
90% and a specificity of between 72% and 100%.5,6 Wolf
et al. studied a total of 220 patients with tonsillitis and
glandular fever and concluded that the lymphocyte to
white cell count ratio should be used as a screening test to
decide on further Monospot testing.6 A lymphocyte to
white count ratio value of 0.35 was found to be the most
sensitive and specific indicator for the presence of EBVrelated IM, with a specificity of 100% and sensitivity of
90%.6 However, Lennon et al. produced a higher-powered
study disputing the accuracy of this method.5 They studied 1,000 patients, again using the same lymphocyte to
white count ratio (0.35), but revealing a reduction in the
specificity (72%) and sensitivity (84%) with a larger
cohort.5 They concluded that the use of the lymphocyte to
white count ratio was neither sensitive nor specific
enough to fully diagnose or exclude EBV-related IM, and
recommended that practitioners revert back to the use of
the Monospot test to accurately ascertain the presence of
EBV-related IM, and only use the lymphocyte to white
cell count ratio to aid in the identification of possible
causes while awaiting formal serological testing (i.e.,
Monospot or EBV serology).5
When examining the results of the age group subanalysis, it can be seen that in the pediatric population
the lymphocyte count had lower positive/negative predictive values, sensitivity, and specificity compared to the
population aged >18 years. The reported sensitivity of
the Monospot test in the pediatric population can be as
low at 50%; therefore, this test should be used with caution in this group, particularly in those younger than 14
years, where a lymphocyte response may be limited or
even absent.2,4 This response may have been a factor
associated with a drop in the predictive ability of the
lymphocyte count seen in the results of our 18-year-old
population. A similar scenario may also occur with atypical presentations of IM, including the immunocompromised, the elderly, and those with neurological
complications, where use of the lymphocyte count alone
is likely to be insufficient to exclude the diagnosis. Many
TABLE IV.
Predictive Value of Lymphocyte Count in Relation to Monospot Results With Age-Specific Subanalysis.
Monospot Positive
Monospot Negative
Population
Lymphocyte
Count > 4 3 109/L
Lymphocyte
Count 4 3 109/L
Lymphocyte
Count > 4 3 109/L
Lymphocyte
Count 4 3 109/L
Positive
Predictive
Value
Negative
Predictive
Value
Sensitivity
Specificity
Whole population
42
8
38
638
53%
99%
84%
94%
18 years old
15
8
22
167
41%
95%
65%
88%
18 years old
27
0
17
470
61%
100%
100%
97%
Laryngoscope 123: October 2013
Biggs et al: Lymphocyte Count as Diagnostic Screen
2403
of these individuals may have false-negative results on
Monospot testing due to the failure of heterophile antibody formation. In the above patient groups, including
the pediatric population, EBV serology is advocated as
the investigation of choice.
Study Strengths
This is a highly powered study (>90%). The sensitivity results from this study are similar to those of Lennon et al. (84%).5 However, when examining the results
of the age subanalysis, those patients older than 18
years have an associated sensitivity of 100%, significantly superior to all other reported studies within the
literature. The negative predictive value of this study is
also significantly higher (99% in all patients studied, rising to 100% in those aged >18 years) than those of similarly powered studies in the use of the lymphocyte to
white cell count ratio (82%).5 However, the unique feature of this study is the novel use of the lymphocyte
count as the sole predictor for a negative Monospot test.
This is significantly quicker than Monospot testing and
does not require any ratio calculations, making it much
easier to interpret and use in clinical practice.
Study Limitations
Within the study, the presence or exclusion of EBV
infection was correlated with a positive or negative Monospot result, respectively. Monospot tests, including the
one used in this study (Mono-Latex), lack perfect sensitivity and specificity, and therefore there is likely to be a
minor degree of inaccuracy associated with their use that
may have influenced the study outcomes. However, due
to cost implications, it would have been impractical to use
EBV serology (the gold standard) in all cases.
Implications for Practice
Severe sore throat and/or tonsillitis are extremely
common. Results of a community postal survey performed in Scotland in 2005 revealed that 29.4% (4,646 of
15,788) of those questioned had suffered tonsillitis or a
severe sore throat in the preceding year, with 38.4%
(1,782 of 4,646) presenting to their doctor with these
symptoms, and 2.1% subsequently referred to a hospital.7 EBV-related IM is important to differentiate due to
its increased associated complications and to enable
appropriate counseling in the recovery period.2,8–10 The
only highly sensitive and specific way of currently identifying these patients is to undertake Monospot or EBV
serological testing. This study introduces a highly sensitive and specific alternative—the lymphocyte count—
which is significantly cheaper, quicker to undertake, and
easier to interpret.
Within the study period, use of the lymphocyte
count as a marker of a negative Monospot test could
have prevented Monospot testing in 646 patients. The
current cost of a Monospot test is £1.29 per test, indicating a potential saving of £833 over the study period (15
weeks). A FBC is significantly cheaper at £0.50 per test
(within this study, 726 of 749 patients had an FBC perLaryngoscope 123: October 2013
2404
formed in addition to a Monospot anyway), with EBV serology significantly more expensive (EBV serology capsid
immunoglobulin G [IgG] 5 £6.75 per test, EBV serology
nuclear antigen IgG 5 £9.68 per test, and EBV serology
polymerase chain reaction 5 £24.39 per test). These figures would translate into significant savings, which is
particularly relevant in the current economic climate
and with health services worldwide facing intense financial pressures.
CONCLUSION
Monospot testing to exclude IM in patients presenting with sore throat is common. This study shows that a
lymphocyte count of 4 3 109/L appears to be a highly
reliable predictor of a negative Monospot result in the
adult population. Thus, rather than employing a blanket
approach, Monospot testing should be reserved for those
adult patients with a lymphocyte count of >4 3 109/L or
for those in whom the clinical symptoms and signs are
highly suggestive of IM. If a Monospot test is required
following FBC analysis, then the laboratory can be contacted and the test requested on the stored FBC sample,
thus negating any need for retesting the patient. This
novel use of the lymphocyte count could dramatically
reduce the need for routine Monospot testing and its
associated costs in a large cohort of patients.
Ethical Approval
This was an anonymized, retrospective case note
and blood results review. Therefore, ethical approval
was not required.
Acknowledgment
The authors thank Mr. Neil Gillett and the Haematology
Laboratory at UHS for their help with this study, and Mr.
Brian Yuen for his valuable assistance with the statistical
analysis.
BIBLIOGRAPHY
1. Graser F. 100 years of Pfeiffer’s glandular fever. Klin Padiatr
1991;203:187–190.
2. Papesch M, Watkins R. Epstein-Barr virus infectious mononucleosis. Clin
Otolaryngol Allied Sci 2001;26:3–8.
3. Hoagland RJ. Infectious mononucleosis. Prim Care 1975;2:295–307.
4. Elgh F, Linderholm M. Evaluation of six commercially available kits using
purified heterophile antigen for the rapid diagnosis of infectious mononucleosis compared with Epstein-Barr virus-specific serology. Clin Diagn
Virol 1996;7:17–21.
5. Lennon P, O’Neill JP, Fenton JE, O’Dwyer T. Challenging the use of the
lymphocyte to white cell count ratio in the diagnosis of infectious mononucleosis by analysis of a large cohort of monospot test results. Clin Otolaryngol 2010;35:397–401.
6. Wolf DM, Friedrichs I, Toma AG. Lymphocyte-white blood cell count ratio:
a quickly available screening tool to differentiate acute purulent tonsillitis from glandular fever. Arch Otolaryngol Head Neck Surg 2007;
133:61–64.
7. Hannaford PC, Simpson JA, Bisset AF, Davis A, McKerrow W, Mills R.
The prevalence of ear, nose and throat problems in the community:
results from a national cross-sectional postal survey in Scotland. Fam
Pract 2005;22:227–233.
8. Macsween KF, Crawford DH. Epstein-Barr virus-recent advances. Lancet
Infect Dis 2003;3:131–140.
9. O’Connor TE, Skinner LJ, Kiely P, Fenton JE. Return to contact sports following infectious mononucleosis: the role of serial ultrasonography. Ear
Nose Throat J 2011;90:E21–E24.
10. Vine LJ, Shepherd K, Hunter JG, et al. Characteristics of Epstein-Barr virus hepatitis among patients with jaundice or acute hepatitis. Aliment
Pharmacol Ther 2012;36:16–21.
Biggs et al: Lymphocyte Count as Diagnostic Screen