False-positive troponin I elevation in a newborn with neonatal encephalopathy

Abstract

Serum biomarkers of myocardial damage are commonly used in babies after perinatal asphyxia. We present a case report of a persistently troponin I elevation without evidence of clinical or instrumental signs of myocardial ischaemia in a baby with perinatal asphyxia. When the blood was mixed with polyethylene glycol we found that the troponin I levels were falsely elevated due to interfering antibodies. This case shows that analytical errors may still occur despite modern immunoassay systems and underlines the need for further investigations to identify false-positive values in case of disagreement between clinical conditions and laboratory values.

Background

Cardiac troponins are biochemical markers of myocardial damage, and their levels increase in newborns with perinatal asphyxia.¹ Measurement of troponins has replaced the creatine kinase-MB isoform as the biochemical test for cardiac involvement. However, with the widespread use of high-sensitivity troponin assays, reports of false-positive or false-negative results have come out in the adult population.^2–4

Troponin levels have been shown to have a good prognostic value with regards to outcome in perinatal asphyxia, with high cardiac troponin I levels correlating well with neurodevelopment at 18 months.⁵ Given its use as biomarker, a possible false-positive result can have important implications from a clinical point of view and can mislead clinicians and lead to unnecessary treatments and interventions.

Case presentation

A baby boy (40 weeks; 3300 g, 25th centile) was delivered following a uneventful pregnancy by ventouse delivery because of failure to progress in the second stage of labour. The baby was born pale, floppy with no respiratory effort and heart rate requiring intubation and ventilation soon after birth. Apgar scores were 0, 5 and 6 at 1, 5 and 10 min, respectively; arterial cord pH showed acidosis with a pH 7.01, lactate 6 and base excess −13.

He was subsequently admitted to Neonatal Intensive Care Unit (NICU) with a diagnosis of hypoxic ischaemia injury. At admission, the baby was hyperalert, with mild flexion of distal joints and weak suck. Amplitude-integrated electroencephalography was normal. The neurological examination was in keeping with mild encephalopathy and therefore the baby did not fulfil the criteria for therapeutic hypothermia.

The patient was extubated after 12 hours since the admission and inotropic support was never required. Renal function was always normal and bilirubin values were always below the phototherapy threshold.

Investigations

Cardiac troponin I measurement (Dimension vista 500, SIEMENS) revealed a high level of cardiac troponin I: 0.58 µg/L (normal <0.045). This is a sandwich chemiluminescent immunoassay, which includes two synthetic bead reagents and a biotinylated anticardiac troponin I monoclonal antibody fragment. The assay range is between 0.015 and 40 µg/L. Concentrations of creatine kinase-MB (6.3 ng/mL) and myoglobin (13.2 ng/mL) were within normal ranges. Other samples were sent after 2, 4, 7 and 15 days since the admission to the unit and confirmed high troponin I levels (0.55, 0.59, 0.51 and 0.56 µg/L, respectively).

The patient had electrocardiography and echocardiogram performed daily to exclude the presence of any myocardial damage. Electrocardiography showed no ischaemic changes, whereas echocardiogram showed good bilateral ventricular function with normal anatomy.

In view of the disagreement between the clinical condition and troponin I levels, different steps were taken. Results were repeatedly confirmed by careful selection of test tubes; both lithium–heparin plasma and serum were tested and yielded similar results. Next, we performed a dilution test to exclude false-positive results. However, these results were inconclusive. Patient’s blood was mixed with polyethylene glycol at a 1:1 ratio to precipitate any antibodies that might be interfering with the cardiac troponin I assay. The sample was cooled, centrifuged and then the supernatant was removed for analysis. Cardiac troponin I levels went down to 0.02 µg/L, suggesting the presence of antibodies, which were interfering with the troponin assay.

Differential diagnosis

Differential diagnoses included cardiac versus non-cardiac origins of troponin I elevation (Table 1). Myocardial injury is frequent in perinatal asphyxia.⁶ Therefore, we first ruled out cardiac causes basing on the normality of echocardiography, electrocardiography and the other myocardial injury biomarkers. We then considered non-cardiac causes of troponin I elevation. These latter, however, were considered unlikely because of the normality of bilirubin and creatinine levels.

Table 1.

Possible causes and investigations to exclude or confirm troponin I elevation

Causes	Investigations
Myocardial injury Myocardial ischaemia Congestive heart failure Supraventricular tachycardia with haemodynamic compromise	Electrocardiography Echocardiogram Creatine kinase-MB, myoglobin levels
Renal failure	Serum creatinine
Hyperbilirubinemia	Serum bilirubin
Analyser malfunction	Repeat measurement with different type of tube (lithium–heparin plasma and serum) Use of another analytical system
Heterophile antibodies (human antianimal antibodies, human antimouse antibody) and autoantibodies	Dilution Precipitation with polyethylene glycol Antibody blocking agents
Presence of residual fibrin or other microparticles in specimen	Heparinising the tubes before analysis Repeating centrifugation to remove excess fibrin Adding a clot activator (ie, thrombin)

We subsequently investigated the possibility of falsely elevated cardiac troponin I levels. We performed a review of the literature. However, there were no previous reports of a falsely elevated cardiac troponin I result in neonates. Therefore, we followed a previous published algorithm for the detection of false-positive troponin measurements in adults.⁷

We excluded potential reasons for preanalytical errors and repeated the blood sample collection ensuring that this was not haemolised. We also considered blood sample processing issues like incomplete separation of serum and presence of residual fibrin or other microparticles. We, therefore, inspected the specimen for the presence of fibrin clots and other microparticles and repeated the centrifugation of the sample before the analysis. Cardiac troponin I levels were also measured in both lithium–heparin/EDTA plasma and serum with similar results.

We then considered analytical errors. To confirm the analyser function, results were repeatedly measured and confirmed in another laboratory of the same hospital. Although there are many reports in adults showing the high prevalence of interfering antibodies (up to 40%),^{8 9} there is a paucity of data in newborns. The presence of heterophilic antibodies has been described as a cause of falsely elevated thyroid stimulating hormone (TSH) values in TSH radioimmunoassay systems commonly used in neonatal screening programmes. In these case series, the authors suggested that neonatal heterophilic antibodies might be due to maternal transfer across the placenta of antibodies of IgG classes.^{10 11} Other possible causes of circulating antibodies include immunotherapies and blood transfusions.⁷ However, no blood transfusion was given.

Outcome and follow-up

Outpatient evaluation 1 month later showed no symptoms with repeat cardiac troponin I at 0.02 ng/mL. The electrocardiography and echocardiogram in this occasion were normal.

Discussion

The introduction of high-sensitivity cardiac troponin assays allows measurement of extremely low levels of troponin with excellent precision and has certainly improved the sensitivity of detection of myocardial ischaemic insults.

Different reports in adults have shown that false-positive troponin I elevation is possible, especially as consequence of increased concentrations of autoantibodies and heterophile antibodies.^{2 3} However, there is no previous report of heterophile antibodies as cause of cardiac troponin I false elevation in the neonatal population.

Endogenous antibodies can be either heterophile antibodies, human antianimal antibodies or autoantibodies and all of them can interfere with immunoassays. Heterophile antibodies are natural antibodies produced against poorly defined antigens. Human antianimal antibodies instead, are considered species specific and produced after acute or chronic exposure to the animal immunoglobulins.

Previous reports highlighted that heterophilic antibodies might be the cause of a factitious neonatal hyperthyrotropinemia by interfering with THS radioimmunoassay systems.¹⁰ However, endogenous antibodies can potentially interfere with any immunoassays including D-dimer¹² and multiple hormone immunoassays as recently highlighted by García-González et al.⁹

Presence of heterophilic antibodies is a well-known cause of interference in high-sensitivity cardiac troponin I immunoassays. These antibodies, in fact, may crosslink with the capture of antibodies during the assay. Currently, many manufactures have developed sandwich-type immunoassay based on chimeric mouse–human antigen-binding antibodies and added blocking agents to limit the presence of a false-positive signal. However, this is still possible especially in case of a large amount of heterophilic antibodies as highlighted by recently published case reports.^13–15

Giacchi et al reported a case of falsely high troponin I levels in a baby with jaundice. However, in our case, bilirubin levels had always been within the normal ranges and the baby never required phototherapy.¹⁶

Previous studies pointed out that there is significant difference in high-sensitivity cardiac troponin I concentrations when measured in lithium heparin versus EDTA plasma with high-sensitivity cardiac troponin I concentrations higher in lithium–heparin versus EDTA plasma.¹⁷ We therefore, tested both lithium–heparin/EDTA plasma and serum with similar results.

In our neonatal unit, the median length of stay in case of perinatal asphyxia with no need for therapeutic hypothermia was 4 days in the last year. However, in this case report, infant’s hospitalisation was prolonged considerably because of the uncertainty of the diagnosis, which led to repeated blood samples and several cardiology referrals. This highlights that clinicians must be aware of the potential analytical errors of modern immunoassays to avoid unnecessary investigations and days of hospitalisation to the patients.

The case presented suggested that a false elevation of troponin levels is possible also in the neonatal population. However, caution is needed before drawing any conclusions because this is a single case report and these data will need to be verified further in other studies with a longer duration of clinical follow-up.

Learning points.

• False-positive results are still possible with modern immunoassays.

• In case of disagreement between clinical conditions and troponin I levels, different steps must be taken to exclude a false elevation of the troponin I levels.

• The presence of interfering antibodies is a possible cause of false-positive troponin I elevation and needs to be excluded in neonates.