Variation of High-Sensitivity Troponin T Results in Patients Undergoing Continuous Renal Replacement Therapy

Abstract

Background:

Cardiac troponins are the preferred biomarker to diagnose myocardial injury. Complicating the interpretation of serial troponins in patients with end-stage renal disease, it has been shown that the hemodialysis procedure results in a small but significant decline in high-sensitivity cardiac troponins (hs-cTnT). This raises the possibility that continuous renal replacement therapy (CRRT) might similarly alter cardiac troponin levels and affect their interpretation when cardiac ischemia is being considered.

Objective:

We sought to determine the effect of CRRT on hs-cTnT levels over time in a group of patients without active myocardial injury.

Design:

Prospective, observational study

Setting:

Single tertiary care hospital, Montreal, Quebec

Patients:

Ten critically ill patients with acute kidney injury (AKI) undergoing CRRT. Cardiac ICU (intensive care unit) patients and acute coronary syndrome patients were excluded from the study. The CRRT prescription was at the discretion of the treating intensivist and relatively high doses were used in this study.

Measurements:

The hs-cTnT levels were drawn pre-CRRT, within 6 hours of initiation, and approximately every 6 hours thereafter along with routine CRRT blood work.

Methods:

Changes in hs-cTnT, creatinine, and albumin levels were recorded over the course of CRRT. Mean change in serum analyte concentration and 95% confidence interval was determined for specified time intervals relative to baseline, with paired t tests used to determine statistical significance.

Results:

Among the 10 patients included in the study, the cause of AKI was primarily acute tubular necrosis from septic shock or hemorrhagic shock. Compared with baseline hs-cTnT levels prior to CRRT initiation, mean hs-cTnT level fell by 42% at 5 to 10 hours post-CRRT initiation, followed by a plateauing of levels for the duration of time on CRRT.

Limitations:

Single-center study only applicable to hs-cTnT assay.

Conclusions:

This study demonstrates a significant decrease in hs-cTnT within 5 to 10 hours of CRRT initiation. This suggests that interpretation of cardiac troponin changes during CRRT must take into consideration the timing of dialysis initiation relative to the time of sample collection.

What was known before

In the hemodialysis (HD) population, it has been shown that high-sensitivity cardiac troponin (hs-cTnT) falls during and shortly after HD, before rising back to baseline levels.

What this adds

This study demonstrates a larger drop in high-sensitivity cardiac troponin (hs-cTnT) levels during CRRT, which must be taken into consideration when interpreting cardiac troponin changes in this setting. Further studies are required to see whether these findings extend to different troponin assays and in patients with established myocardial injury.

Introduction

Acute kidney injury (AKI) is common among patients admitted to the intensive care unit (ICU), with 57% of such patients meeting the diagnostic criteria for AKI. ¹ Among patients admitted to the ICU with AKI, dialysis may be required, and continuous renal replacement therapy (CRRT) is often the modality of choice when hemodynamic instability is an issue. Cardiovascular disease is a significant contributor to AKI-associated mortality. ² Diagnosing an acute coronary syndrome (ACS) in this setting can be challenging as patients may present with nonspecific symptoms, and as many as 25% of patients can have silent ischemia. ³ Furthermore, critically ill patients are often sedated and intubated. As such, cardiac biomarkers are relied upon to diagnose myocardial injury. In addition, interpretation of cardiac troponins (cTns) can be more difficult in patients with renal disease. For example, patients with chronic kidney disease (CKD) and AKI have been noted to have stable but elevated cTn levels. ⁴ Such cTn elevations have been demonstrated to have long-term prognostic implications for cardiovascular mortality, but are not necessarily a marker of acute myocardial injury. ⁵ As such, dynamic changes in cTn remain key to diagnosing an ACS, both in the general population and in patients with AKI and CKD. The impact of renal replacement therapy (RRT) on troponin levels poses another challenge. The molecular weights of cardiac troponin I (cTnI) and cardiac troponin T (cTnT) are 24 kDa and 36 kDa, respectively. The pore size of filters used for CRRT is typically approximately 30 kDa. As such, one would expect relatively little clearance of these molecules. Data in the chronic hemodialysis (HD) population have shown variable results. One study employing non-high-sensitivity cTn demonstrated cTn stability or rise following HD. ⁶ Another study using high-sensitivity cTnT (hs-cTnT) demonstrated no reduction after conventional HD, but interestingly a significant reduction in serum levels after hemodiafiltration, particularly with a longer duration. ⁷ In a study of 10 chronic HD patients, we recently demonstrated that hs-cTnT falls by approximately 11% during and shortly after HD, before rising back to baseline levels 11 hours later. ⁸ The purpose of this study was to describe the influence of CRRT on hs-cTnT levels among patients admitted to the ICU with AKI.

Materials and Methods

This was a single-center, prospective observational study conducted at the Jewish General Hospital (Montreal, Quebec, Canada). Patients included were those admitted to the ICU with AKI undergoing CRRT (initiated at the discretion of the attending intensivist). Exclusion criteria included patients with ACS, post-cardiac surgery patients, and those with end-stage renal disease (ESRD). Patient demographics, reason for ICU admission, reason for AKI, and outcome were recorded. All measures of hs-cTnT were blinded to the treating team. Five patients were on continuous veno-venous hemodiafiltration, one patient was on continuous veno-venous hemodialysis, and 4 patients were on continuous veno-venous hemofiltration. The CRRT parameters are included in Table 1. Prismaflex ST150 dialyzers (Baxter Healthcare, Deerfield, Illinois) were used in all patients, and CRRT prescription and dosing was at the discretion of the treating intensivist. The study received ethics approval from the hospital’s research ethics board and was adherent to the Declaration of Helsinki.

Table 1.

Patient Characteristics.

Patient	Age (y)	Sex	Patient weight (kg)	Cause of renal failure	Reason for ICU admission	Duration of CRRT (h)	Creatinine at CRRT initiation (µmol/L)	Urine output (cc/24 h)	Dialysate rate (cc/kg/h)	Replacement rate a (cc/kg/h)	Ultra-filtration (cc/h)	Anticoagulation	Outcome
1	67	M	85	ATN	Intra-abdominal sepsis	41	410	<200	23.5	23.5	0	Heparin (800-1000 U/h)	Renal recovery
2	59	F	61.5	ATN	Pneumosepsis with multiorgan dysfunction	25	228	<200	None	32.5	0	None	Died
3	58	M	196	ATN	Sepsis and hypercapnic respiratory failure	38	593	<200	10.2	None	0	Heparin (800-1400 U/h)	IHD
4	49	F	88	ATN	Post-surgical sepsis and hemorrhage	66	555	<200	22.7	22.7	0	None	Died
5	68	M	111	ATN	Post-surgical sepsis	56	246	<200	18	18	0	None	IHD
6	86	M	99	ATN	Ascending cholangitis	27	668	<200	15	15	50	Heparin (400-500 U/h)	Died
7	77	F	108.5	ATN	Urosepsis	63.5	558	<200	None	27.6	25	None	IHD
8	36	F	82.5	ATN	Eclampsia and hemorrhagic shock	65	401	<200	None	24.2	100	None	IHD
9	61	M	82	ATN	Sepsis	61	288	450	24.4	24.4	100	None	Died
10	29	M	127	ATN	Hemophagocytic lymphohistiocytosis and lymphoma	72	600	<200	None	7.9	50	None	Died

Note. ICU = intensive care unit; CRRT = continuous renal replacement therapy; ATN = acute tubular necrosis; IHD = intermittent hemodialysis; CKD = chronic kidney disease.

a

Pre-filter replacement with 200 cc/h post-filter replacement.

Baseline hs-cTnT was measured prior to initiation of CRRT (mean of 3.9 hours prior to initiation with SD of 3.2), within 6 hours of initiation, and subsequently, approximately every 6 hours. The exact frequency of blood collection was at the discretion of ICU nursing staff based on routine clinical care. Given the variable duration on CRRT, patients had varying number of measurements. The hs-cTnT was determined using a fifth-generation hs-cTnT immunoassay with a coefficient of variance <10% at the 99th percentile (Roche Diagnostics, cat no. 05092744). Creatinine and albumin were also measured with each collection (Roche Diagnostics, cat no. 03263991190 and 11815148216).

To standardize time of collection, analytes were grouped into time intervals: 5 to 10 hours, 11 to 15 hours, 16 to 20 hours, and 21 to 30 hours post-initiation of CRRT. Mean change in analyte concentration and 95% confidence interval was determined at each time interval relative to baseline. Paired t tests were used to determine statistical significance of within-patient differences over time (Microsoft Excel).

Results

Ten patients were enrolled in this study (6 males and 4 females) (Table 1). Median age was 59 years (range = 29-86). All patients were admitted to the ICU with shock, with sepsis being the etiology in 80%. Mean duration of CRRT in this cohort was 51.5 hours with a range of 25 to 72 hours. Mean creatinine prior to CRRT initiation was 454.7 µmol/L (SD of 161.3). Five patients died, 1 had sufficient renal recovery to discontinue RRT, and 4 patients were transitioned to intermittent HD post-ICU discharge.

There was a significantly lower mean hs-cTnT during CRRT in each time interval, when compared with levels prior to the start of dialysis (P < .05) (Figure 1). There was a 42% decrease in hs-cTnT at 5 to 10 hours post-CRRT initiation, with stable levels thereafter (Figure 1). As expected, creatinine steadily declined after CRRT initiation in all 4 time intervals (P < .05), while there was no significant change in albumin levels during CRRT. In the 3 patients in whom hs-cTnT was measured after CRRT discontinuation (5-10 hours after last blood work drawn during CRRT), a rise in hs-cTnT was noted (P = .057).

Figure 1.

Change in analyte levels and corresponding confidence intervals over the course of CRRT.

Discussion

This study has demonstrated significant clearance of hs-cTnT in patients undergoing CRRT. By 5 to 10 hours post-initiation of CRRT, hs-cTnT levels decreased by 42% from baseline and remained stable thereafter. In patients with available blood work, there appeared to be a rise in hs-cTnT after cessation of CRRT, although this did not reach statistical significance.

Stably elevated troponin levels have been demonstrated in patients with CKD and AKI, likely reflecting subclinical coronary artery disease. ⁴ As such, dynamic changes in cTn are important to diagnose ACS. This study, in conjunction with our previous investigation in HD patients, offers some insight into the influence of dialysis on cTn levels. While an 11% reduction in mean hs-cTnT was demonstrated during HD, ⁸ we noted a much larger reduction in CRRT patients. This likely reflects the longer duration of dialysis and possibly the greater convective clearance of larger molecular weight substances such as hs-cTnT (and/or their degradation products, which have been shown to cross-react with the hs-cTnT assay). ⁹ This is also consistent with the study by Cardinaels and colleagues ⁷ demonstrating greater reduction in serum levels of hs-cTnT with longer duration of hemodiafiltration.

Clinical guidelines currently recommend using a 20% or more change in cTn levels from baseline to diagnose ACS. ¹⁰ Thus, the greater than 40% reduction in troponin observed in this study can mask an ACS. Alternatively, a lack of fall in troponin levels upon CRRT initiation should be concerning for myocardial injury. It is important to note that a steady state appears to be reached 5 to 10 hours after CRRT initiation, after which a new baseline hs-cTnT is reached, assuming no further interruptions in CRRT. One might therefore infer that myocardial injury occurring 5 to 10 hours after starting CRRT should manifest as a rise in hs-cTnT above this new baseline.

To our knowledge, this is the first study showing a significant reduction in cTns after CRRT initiation, in keeping with the hs-cTnT studies in HD patients. Nevertheless, there are several study limitations. First, this was a single-center study with a small sample size and a lack of homogeneity in CRRT prescriptions. Second, only hs-cTnT was studied, so the results cannot be extrapolated to non-high-sensitivity troponin assays or to Troponin I assays. Third, we cannot say with certainty whether the reduction was due to clearance of TnT or of its metabolites which can interfere with the assay. It would be interesting to measure effluent concentrations of hs-cTnT and its metabolites in future studies to explore this further. Fourth, we cannot rule out the possibility of subclinical cardiac ischemia in some patients prior to initiation of CRRT, which could have partly explained the declining hs-cTnT observed after CRRT initiation. Finally, the majority of patients received fairly high doses of CRRT, which likely influenced the magnitude of hs-cTnT removal and must be taken into consideration when anticipating changes in levels. Despite these limitations, our study is novel and sheds light on the highly clinically relevant question of what happens to hs-cTnT levels in patients on CRRT. Further studies are warranted to confirm our findings, and to determine whether this effect is seen with different troponin assays and how this affects post-ACS and post-cardiac surgery patients.

To conclude, hs-cTnT levels decrease within 5 to 10 hours of CRRT initiation. Consequently, interpretation of troponin changes during CRRT must take into consideration the timing of CRRT initiation relative to the timing of sample collection.