Lipoprotein(a) May Be Falsely Lowered When Troponin T is Elevated in Patients with Ischaemic Heart Disease During Acute Admissions

Ischaemic heart disease (IHD) and its sequelae remains a leading cause of morbidity and mortality worldwide. Lipoprotein(a) – Lp(a) – has been shown to increase the risk of atherosclerotic events in epidemiological and Mendelian randomisation studies.¹ Lp(a) consists of an apolipoprotein B (apoB) and an apolipoprotein A moiety causing prothrombotic, proinflammatory and proatherogenic features.² As a result, elevated Lp(a) increases the risk of atherosclerotic events, such as coronary artery disease and acute coronary syndrome (ACS), calcific aortic valve stenosis and cardiovascular mortality.³ Multiple guidelines recommend Lp(a) to be measured at least once in an adult’s lifetime to allow early intervention to prevent atherosclerotic cardiovascular disease.^4,5 Elevated Lp(a) is under-diagnosed worldwide and measuring Lp(a) during hospital admissions would be an effective opportunistic early detection strategy.⁶ Lp(a) level is predominantly genetically controlled and the non-genetic variability of Lp(a) is thought to be less than 20%.⁷ It remains inconclusive whether Lp(a) levels are affected during ACS.^8,9

We sought to investigate if Lp(a) concentration correlates with elevated troponin T concentrations in patients admitted to hospital with IHD.

Methods

We analysed a subgroup of a prospective cohort of a multiethnic Asian population with IHD during admission to a tertiary hospital in Singapore.¹⁰ The original study was a prospective cohort of 555 patients admitted to a single tertiary centre for IHD. The primary outcome of the original study was to assess the prevalence of elevated Lp(a) in patients diagnosed with IHD. The study received ethics approval from the Singhealth institutional review board. All the patients provided informed consent to participate in the study. Inclusion criteria for the original study were that candidates must be able to give informed consent and had been admitted to hospital with a primary diagnosis of IHD. Exclusion criteria were those who declined to give informed consent.

The primary objective of this subgroup analysis was to determine if the baseline troponin T is associated with a change of Lp(a) on follow up. The secondary objective was to explore if the association of troponin T and Lp(a) is modified by acute phase inflammation. This subgroup analysis from the original study included only patients with baseline Lp(a) ≥70 nmol/l and at least one repeat Lp(a) concentrations outpatient when clinically well. All patients had either two or three serum Lp(a) concentrations, over a median of 13 months. Changes in serum Lp(a) concentration were stratified into ≥20% decrease, <20% change and ≥20% increase of Lp(a) concentration at follow-up for multivariable proportional ORs from logistic regression analysis. Multivariable proportional odds logistic regression was used to identify independent explanatory variables of a change in age, gender, BMI, renal impairment and new statin use.¹⁰ Analyses were repeated with ordinal coding using 10%, 25%, or 30% threshold.

A clinical diagnosis of elevated Lp(a), otherwise called hyper-Lp(a) was defined as Lp(a) ≥120 nmol/l.³ Lp(a) concentrations within the grey zone (i.e. borderline hyper-Lp(a)) were defined as ≥70 and <120 nmol/l. Serum Lp(a) was measured using particle-enhanced turbidimetric immunoassay with Tina-quant lipoprotein(a) Gen.2 (Roche) and the inter-assay coefficient of variation (CV) were less than 2.2. High-sensitivity serum tests were made for troponin T, C-reactive protein (CRP) and serum interleukin-6 (IL-6) using the Cobas 8,000 modular analyser.

Descriptive statistics for the patients’ baseline demographic and clinical characteristics were reported as number and percentage for categorical data, mean ±SD for normally distributed data, and median and interquartile range (IQR) for other data. Multivariable proportional odds logistic regression was used to identify independent explanatory variables of a change in Lp(a) level at follow-up. Covariates included in the model were selected based on the literature and expert clinical opinion. Unadjusted and adjusted proportional OR with corresponding 95% confidence intervals were reported. The proportional odds assumption, which posits that the effect of an explanatory variable is constant for each increase in the level of the outcome, was assessed using Brant’s test.

To explore if there was effect modification by acute phase inflammatory responses, we performed subgroup analyses of serum CRP and IL-6 concentrations. We included an interaction term of CRP or IL-6, and troponin T in the multivariable model to check for effect modification on the multiplicative scale. All statistical analyses were conducted using Stata 18 (StataCorp).

Results

A total of 75 patients were identified as having Lp(a) ≥70 nmol/l and had at least one or two Lp(a) measured during follow-up. The mean age at baseline was 62.2 (±11.2) years and 20% (n=15) of the patients were women. The median follow-up time was 13 months (IQR 8–14). Patient demographics and clinical characteristics are shown in Table 1. None of the patients had hypothyroidism, active cancer, were consuming steroids or required dialysis and 14.7% (n=11) were newly started on statin because of newly diagnosed IHD during the same hospital admission. None of the patients were on agents that inhibit PCSK9 enzyme activity. 22.7% (n=17) of the patients had mild renal impairment defined by glomerular filtration rate <60 ml/min and none had end-stage renal failure. Over half (57.3%; n=43) of patients were admitted to hospital for an acute MI (AMI).

The median Lp(a) concentration at baseline and final follow-up was 143.2 nmol/l (IQR 90.9–186.0) and 141.4 nmol/l (IQR 76.2–197.0) respectively; p=0.049. The univariate proportional odds logistic regression showed that for a 1 unit increase in log_e-troponin T, equivalent to 2.72 times of troponin T, there was a 29% increase in odds of ≥10% higher Lp(a) concentration at follow-up (OR 1.29; 95% CI [1.03–1.61]; p=0.029) (Table 2). Multivariable analysis showed a 50% increase in odds of ≥10% higher Lp(a) concentration at follow-up (OR 1.50; 95% CI [1.15–1.96]; p=0.003). Similar results were obtained when the ordinal outcome was coded using a 20%, 25%, or 30% threshold. With multivariable regression analysis, higher serum troponin T concentration was associated with ≥20 and ≥30% higher Lp(a) concentration at follow-up with OR 1.34; 95% CI [1.04–1.74]; p=0.025 and OR 1.31; 95% CI [1.01–1.70]; p=0.045, respectively. Brant’s test showed that the proportional odds assumption was violated for new statin users (p=0.001), but not for troponin T (p=0.765), age (p=0.591), sex (p=0.257), BMI (p=0.697), renal impairment (p=0.447), or creatinine (p=0.981). When excluding the new use of statins as a covariate in the multivariable analysis, consistent results were observed. To investigate the effect modification of acute inflammation at baseline (hospital admission) on the relationship between baseline serum troponin T and subsequent changes in serum Lp(a) concentrations, the multivariable model was repeated, adjusting for either CRP or IL-6 concentrations. The adjusted proportional OR was attenuated among patients with lower CRP or IL-6 concentrations but amplified among patients with higher serum CRP or IL-6 concentrations.

Figure 1 shows the percentage of participants whose diagnosis of hyper-Lp(a) was reclassified after repeat blood measurements at follow-up. With increasing baseline serum concentrations of troponin T, there was an increased percentage of patients with a reclassified diagnosis of hyper-Lp(a) from non-hyper-Lp(a). Among patients with baseline serum troponin T concentration of <30 ng/l, only 4.3% of patients had a new diagnosis of hyper-Lp(a) at follow up, whereas among patients with baseline troponin T concentration of ≥30 ng/l, 7.7% were reclassified as hyper-Lp(a). Among patients with baseline troponin T ≥100 and ≥500 ng/l, 10% (4/40) and 13.6% (3/22) had reclassification of diagnosis to hyper-Lp(a) respectively. However, there were also patients that had a reclassification from hyper-Lp(a) to non-hyper-Lp(a) with 13% (3/23) of patients with serum troponin T <30 ng/l and 18.2% (4/22) of patients with serum troponin T ≥500 ng/l. This suggests that acute cardiac-related illnesses can cause variations in Lp(a) in both directions.

There were six patients who had an increase in serum Lp(a) concentrations by more than 50% in degree of elevation after being discharged from hospital (range 50.9–184.9%), two of whom had a >100% increase at follow-up.

One patient was a 50-year-old man of Indian ethnicity without any previous medical history who had been admitted to hospital for ST-elevation AMI with peak troponin T concentration of 25,000 ng/l. He underwent urgent percutaneous coronary stent insertion to his left anterior descending artery which was 80% stenosed. At day four, his Lp(a) was 82 nmol/l, LDL cholesterol was 5.17 mmol/l, CRP was 27 mg/l, troponin T 3,250 ng/l without kidney impairment, sepsis or thyroid disorder. He did have transient elevation of liver transaminases that most likely were related to AMI (ALT 115 U/l, AST 574 U/l, but normal levels of bilirubin and albumin. Transaminitis fully resolved by the next month. He was prescribed atorvastatin after his transaminitis resolved. At 6 months, his serum Lp(a) concentration rose to 160.8 nmol/l and at 12 months it was 198.6 nmol/l (an elevation of 142%). There were no discernible secondary causes to explain the initial decrease of Lp(a) level apart from AMI and possibly the subsequent statin initiation that could have partially increased the Lp(a) concentration.

Discussion

To our knowledge, this is the first study that describes the temporal relationship between troponin T and Lp(a) concentration in adults. We suggest that patients with elevated troponin T concentration in hospital may have falsely lowered Lp(a) concentrations, leading to under-diagnosis of hyper-Lp(a). Lp(a) is approximately 90% genetically determined. The non-genetic variation of Lp(a) is thought to be <20%, unless there is severe renal and liver impairment.⁷ However, we report that serum Lp(a) concentrations can vary by >30%, especially when checked at acute admissions and therefore a repeat Lp(a) concentration should be performed, especially when Lp(a) is in the grey zone of Lp(a) 70–125 nmol/l [30–50 mg/dl].⁵

Of important clinical relevance, we also described two patients with markedly falsely decreased Lp(a) concentrations by >100% during acute hospital admissions and repeat concentrations as an outpatient, demonstrating the extensive variability of Lp(a) during acute clinical events. Although the variation of Lp(a) can be increased or decreased when rechecked when the patient is clinically stable, the falsely lowered Lp(a) at acute admissions can lead to underdiagnosis of hyper-Lp(a) or missing the true extent of the elevated Lp(a) concentration which can mean a missed opportunity to reduce the cardiovascular risk of the patient. The European Atherosclerosis Society recommends that Lp(a) concentrations lower than 70–75 nmol/l (30 mg/dl) would not require any repeat testing for life, whereas elevated Lp(a) >120–125 nmol/l (≥50 mg/dl) is an important cardiovascular risk enhancer and requires intensified management of risk factors and cascade testing should be considered.^3,5

Our finding that the variability of Lp(a) being clinically significant at acute admission is supported by an observational study of 1,245 white people who underwent coronary angiography after presenting with ACS, which showed that Lp(a) was increased at 1–3 months after acute admission, suggesting that Lp(a) can be falsely lowered during the initial ACS event.⁹ Similarly, another study of 99 patients with AMI had significantly higher Lp(a) concentrations at 6 months after AMI.¹¹ In another study, subgroup analysis derived from the population of two randomised controlled studies showed that 74 white patients with Lp(a) >75 nmol/l were shown to have a rise in Lp(a) from ACS presentation at 30 and 120 days afterwards.¹² The median increase of Lp(a) was 28% in patients with Lp(a) >75 nmol/l as compared with an increase of 10% in Lp(a) in those with pre-treatment Lp(a) <75 nmol/l. However, older studies conflict with these findings and report that Lp(a) may be transiently increased in the days following AMI.^8,13 Percutaneous coronary intervention (PCI) may also affect Lp(a) results. A study of 141 patients with stable IHD undergoing PCI demonstrated that the rise of Lp(a) after PCI was in 64% of the patients and this correlated with a 36% increase in oxidised phospholipids. Both Lp(a) and oxidised phospholipid concentrations returned to near baseline 6 hours later.¹⁴ Other studies have also challenged the notion that Lp(a) is generally stable from childhood to adulthood. A study of 2,721 children showed that Lp(a) concentrations in children increased with age, with a variation between individuals of up to 70%.¹⁵ In another study involving 11 patients in the outpatient setting, intra-individual coefficient variations of Lp(a) had a 25–60% variation over a median of 11 years.¹⁶

Troponin is a protein complex that occurs in cardiac myocytes and skeletal muscle cells. Of the three subunits, troponin T and I are specific to heart muscle, hence they are used as specific biomarkers for myocardial injury.¹⁷ Beyond ACS and IHD, troponin may also be elevated in subclinical cardiovascular assessment such as cardiac trauma, drug-induced cardiac toxicity. For our study, we used serum troponin T for analysis in our patients. Serum troponin T is also elevated in other non-cardiac conditions such as sepsis and renal failure, which also affects Lp(a) concentrations. In this study, we showed a correlation of the baseline troponin T concentration with changes in Lp(a) concentrations, which was partially explained by acute inflammation (as indicated by raised CRP and IL-6). This relationship is unlikely to be causal but rather the elevated troponin T concentrations may reflect the degree of severity of the underlying illness, which also affects the variability of Lp(a) concentrations.

Limitations of this research were that our study population was small and follow-up troponin T concentrations were not measured in the outpatient setting. Larger studies are required to confirm our findings and to elucidate the mechanism affecting Lp(a) variability in acute and chronic conditions. Nonetheless, our findings suggest the clinical importance of repeat Lp(a) concentration tests when patients with IHD, raised troponin T and marginally elevated Lp(a) are clinically well. Relying on a single measurement of Lp(a) may give these patients and clinicians a false sense of security if their Lp(a) concentrations fall in the borderline zones and the window of opportunity for appropriate risk management will be missed. However, we also found some patients were reclassified from hyper-Lp(a) to non-hyper-Lp(a), suggesting that acute cardiac-related illnesses can cause variations in Lp(a) in both directions.

Conclusion

Patients may have falsely altered Lp(a) concentrations during hospital admission and an elevated serum troponin T concentration may be a marker that prompts the need for repeated measurement of Lp(a). To avoid under-diagnosis of hyper-Lp(a), patients with borderline elevated Lp(a) concentrations, particularly in the indeterminate zone (70–120 nmol/l), should have their Lp(a) measurements repeated in the outpatient setting during follow-up when they are well.