The laboratory diagnosis of infectious mononucleosis (IM) caused by the Epstein-Barr virus (EBV) is usually made on the basis of a combination of a relative lymphocytosis, more than 10% atypical lymphocytes, and the presence of heterophile antibodies.1 More specific serological testing for an acute infection includes the detection of an immunoglobulin M (IgM) antibody to the viral capsid antigen and the absence of Epstein-Barr nuclear antigen antibodies. In clinical practice, heterophile antibody testing (HAT) is often performed at the same time that a complete blood cell (CBC) count is ordered in situations where IM is suspected. The sensitivity of the HAT is dependent on the duration of symptoms and age of the patient. The positivity rate is 75% in the first week of illness and increases to about 90% by the third week.1 The rate is significantly lower in those younger than 12 years (25%-50% positive).2 The HAT positivity also directly increases with the percentage of lymphocytes in the peripheral white blood cell count. In 1 study, the presence of greater than 50% lymphocytes on the CBC count had an 84.5% specificity for having a positive HAT.3 In a report of 220 people with tonsillitis, a ratio of lymphocytes to the total white blood cell count greater than 0.35 was 100% specific and 90% sensitive in diagnosing IM.4 However, a significant number of HAT-positive patients do not have an absolute lymphocytosis.3,5 In practice, most HATs are requested before the CBC count results are known, and many of the patients will not have IM, thus, the overall percentage of positive HAT results is low. A parameter based on the CBC count was sought that would predict the likelihood of a negative HAT. We compared the HAT positivity rate with lymphocyte/neutrophil (L/N) determined by the absolute number of lymphocytes/absolute number of neutrophils on a CBC count done on the same visit.
St Mary's Hospital is a 316-bed McGill University-affiliated teaching hospital that provides laboratory services for a predominantly adult population.
However, the laboratory also performs testing for the family medical center affiliated with the hospital and for more than 30 medical clinics and more than 300 community physicians. The hospital laboratory information system was searched using a Monarch data retrieval tool to find the results of patients who had both a HAT and CBC count performed at the same visit. A total of 5104 results were obtained between the period of June 1997 and November 2005. The HAT was performed either by using Monospot (Oxoid) or Monogen (Biokit, USA) using the recommended manufacturer's procedure. Results that were reported as weakly positive or equivocal were considered as positive. The laboratory reference range of neutrophils is 1.4 to 6.5 × 109/L and of lymphocytes is 1.2 to 3.4 × 109/L. Results of the HAT were compared with results obtained by comparing the ratio of the absolute lymphocyte count to absolute neutrophil count (L/N) (Fig. 1).
A total of 5104 samples were available for analysis of which 501 (9.8%) were HAT positive. The rate of HAT positivity increased significantly when only samples from patients with an L/N ratio greater than or equal to 0.5 were included (Table 1). Using the ratio of greater than or equal to 0.5, the HAT positivity rate rose to 15.7% (455/2888), a relative increase of 60%. A total of 2214 patients had a ratio of less than 0.5, and this represented 43.4% of all the patients tested. The HAT positivity rate in this group was 2.1% and represented 9% of the total positives. The HAT positivity rate increased with increases in the L/N ratio, reaching a maximum of 78.8% at an L/N ratio of 4.1 (Fig. 1). Of the 501 HAT-positive results, 153 (30%) did not demonstrate an absolute lymphocytosis (≥3.5 × 109/L) (Table 2). There were 15 patients with equivocal HAT results, 4 of which had an L/N ratio of less than 0.5. Information on patients' ages was available on 4990 test results (492 positives) and is summarized in Table 3. The age range was 2 to 91 years. Of patients having an L/N ratio of less than 0.5 and a positive HAT, 50% were less than or equal to 25 years old, and this age group accounted for 76.2% of all HAT positives.
Infectious mononucleosis is a syndrome characterized by fever, pharyngitis, adenopathy, and malaise, and is caused by EBV.6 It occurs most frequently in adolescents with the highest rate occurring in 10- to 19-year-old patients. Laboratory confirmation of IM is based on Hoagland: greater than or equal to 50% lymphocytes with at least 10% atypical lymphocytes and a positive HAT.1 Although very specific, these criteria are not very sensitive, and so newer EBV-specific tests have been developed (IgG/IgM antibody to viral capsid antigen and Epstein-Barr nuclear antigen antibodies).7 These tests are considered the diagnostic standard for IM but are expensive and usually not available on a same-day basis, so the HAT remains the most commonly ordered confirmatory test for IM. The major limitation of the HAT is the low sensitivity in patients suspected of having IM. In 1 series, 1921 samples were tested, and only 7.8% of tests were positive.8 The time that a sample is obtained for HAT is important because as many as 25% of cases are negative during the first week of illness.1 This percentage decreases during the next 2 weeks of illness, but 10% of cases of IM will not develop a positive HAT. False-positive results for HAT are believed to be rare but may reflect a recent prior episode of IM (positive results may persist for a year), lymphoma, or a technical error in performing the test. Other acute infectious illnesses (hepatitis, malaria) may also cause a false-positive HAT.
In clinical practice, a HAT is usually ordered concomitantly with a CBC count in patients suspected of having IM. Previous studies showed that a significant number of patients have a positive HAT in the absence of an absolute lymphocytosis.3,5 This study was undertaken to determine if the L/N ratio would be more useful in predicting those patients who were unlikely to have a positive HAT.
In our large sample, the HAT positivity rate was 9.8%. This low value reflects both the limitations of the test and the likelihood that most patients tested do not have IM. It is possible that physicians are ordering HAT to screen for IM. As such, this test performs very poorly. By restricting the HAT to patients having an L/N ratio greater than or equal to 0.5, the positivity rate could be increased to 15.7%, which represents a relative increase of 60%. Only 2.1% of patients with an L/N ratio of less than 0.5 had a positive HAT. Clinical information was not available for most of these patients, so it is unclear whether some of these could be false-positive results. None of these patients had an absolute lymphocytosis. Hoaglund has cautioned clinicians not to rely on the HAT as the sole laboratory parameter to diagnose acute IM.9
Restricting HAT to patients whose L/N ratio is greater than or equal to 0.5 would have reduced the amount of testing by 43.4%, with a 9% (46/501) decrease in the number of positive test results.
There was a progressive increase in the HAT positivity rate with increases in the L/N ratio, with the maximum positivity (78.8%) reached with an L/N ratio of 4.1. Of the 501 positive HAT results in the present study, 153 (30%) did not demonstrate an absolute lymphocytosis (Table 2). Of the group with an L/N ratio greater than or equal to 0.5 and a positive HAT, 23.5% did not have an absolute lymphocytosis. This is similar to findings in other studies where 33.7%3 and 68%5 of HAT-positive patients did not have lymphocytosis. In a tonsillitis study using the ratio of total lymphocytes to the total white blood cell count, 10% of patients with positive HAT did not have an elevated ratio.4 Thus, using absolute lymphocytosis or the ratio of lymphocytes to total white blood cells to restrict testing would miss significant numbers of HAT positives. In the present study, the restriction of HAT to only those with an L/N ratio greater than or equal to 0.5 would have missed only 9% of the positives while reducing the amount of HAT performed. What remains unanswered is how many of the positive HAT patients with L/N ratio of less than 0.5 truly have acute IM. The specificity of the Monogen test is reported as 93% in the product monograph. False-positive results are seen in cases of leukemia, lymphoma, rheumatoid arthritis, viral hepatitis, and cytomegalovirus infection. The HAT can remain positive for months. Specific EBV serologies would need to be performed to determine whether the positive HAT reflected acute IM.
We have implemented a restrictive policy for HAT based on these data. Our current laboratory protocol requires that a CBC count be performed first, and only if the L/N ratio is greater than or equal to 0.5 will the HAT be done, and it can be performed on the tube sent for the CBC count.
In the first year of this policy, the number of HATs performed was reduced from an average of 750 to 235, representing a reduction of 69%. The percentage of HAT positives rose to 29.4%. The greater than expected reduction in testing is probably caused by the requirement of a simultaneous CBC count. On the few occasions where physicians have insisted on performing HAT despite a low L/N ratio, the test has been negative.
This study has limitations that include its retrospective nature, the absence of clinical information, and the lack of atypical lymphocyte enumeration. Testing for EBV-specific antibodies concurrently with HAT testing would have provided information on both false-positive and false-negative HAT. However, physicians should be made aware of the relatively low yield of indiscriminate HAT testing particularly when the L/N ratio is less than 0.5. Laboratories may wish to limit HAT to patients with an L/N ratio greater than or equal to 0.5.
The authors thank Raymond Lussier for data retrieval, and Shelley Leckey, Sunil Joshi, and Michael Gliserman for secretarial assistance.
1. Hoagland RJ. Infectious mononucleosis. Prim Care
2. Schmitz H, Volz D, Krainick-Reichert CH, et al. Acute Epstein-Barr virus infections in children. Med Microbiol Immunol
3. Brigden ML, Au S, Thompson S, et al. Infectious mononucleosis in an outpatient population: diagnostic utility of 2 automated hematology analyzers and the sensitivity and specificity of Hoagland's criteria in heterophile-positive patients. Arch Pathol Lab Med
4. 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
5. Ventura KC, Hudnail SD. Hematologic differences in heterophile-positive and heterophile-negative infectious mononucleosis. Am J Hematol
6. Ebell MH. Epstein-Barr virus mononucleosis. Am Fam Physician
7. Linderholm M, Borman J, Juto P, et al. Comparative evaluation of nine kits for rapid diagnosis of infectious mononucleosis and Epstein-Barr virus-specific serology. J Clin Microbiol
8. Tsaparas YF, Brigden ML, Mathias R, et al. Proportion positive for Epstein-Barr virus, cytomegalovirus, human herpesvirus 6, toxoplasma, and human immunodeficiency virus types 1 and 2 in heterophile-negative patients with an absolute lymphocytosis or an instrument-generated atypical lymphocyte flag. Arch Pathol Med
9. Hoaglund RJ. Diagnosis of infectious mononucleosis. Blood