Within‐individual reproducibility of a dual sample oral cholate challenge test (DuO) and simplified versions of the HepQuant SHUNT test

Abstract

Current noninvasive liver tests measure fibrosis, inflammation, or steatosis and do not measure function. The HepQuant platform of noninvasive tests uniquely assesses both liver function and physiology through the hepatic uptake of stable isotopes of cholate. However, the prototypical HepQuant SHUNT test (SHUNT V1.0) is cumbersome to administer, requiring intravenous and oral administration of cholate and six peripheral venous blood samples over 90 min. To alleviate the burden of test administration, we explored whether an oral only (DuO) version, and other simplified versions, of the test could provide reproducible measurements of liver function. DuO requires only oral dosing and two blood samples over 60 min. The simplified SHUNT test versions were SHUNT V1.1 (oral and IV dosing but four blood samples) and SHUNT V2.0 (oral and IV dosing but only two blood samples over 60 min). In this paper, we describe the reproducibility of DuO and the simplified SHUNT tests relative to that of SHUNT V1.0; equivalency is described in a separate paper. Data from two studies comprising 236 SHUNT tests in 94 subjects were analyzed retrospectively by each method. All simplified methods were highly reproducible across test parameters with intraclass correlation coefficients >0.93 for test parameters Disease Severity Index (DSI) and Hepatic Reserve. DuO and SHUNT V2.0 improved reproducibility in measuring portal‐systemic shunting (SHUNT%). These simplified tests, particularly DuO and SHUNT V2.0, are easier to administer and less invasive, thus, having the potential to be more widely accepted by care providers administering the test and by patients receiving the test.

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Current noninvasive liver tests measure fibrosis, inflammation, and steatosis and do not measure function. The HepQuant SHUNT test assesses both liver function and physiology using naturally occurring endogenous cholates labeled with stable isotopes to probe hepatocyte uptake and hepatic perfusion.

• WHAT QUESTION DID THIS STUDY ADDRESS?

We explored whether the oral‐only HepQuant DuO test and other simplified test versions could provide reproducible measurements of liver function.

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

HepQuant DuO and simplified test versions reduce the number of blood samples by two‐thirds and test time by one‐third and provide accurate measurements of liver function and physiology with high within‐individual reproducibility.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

HepQuant delivers accurate information about hepatic impairment in persons at risk / who have liver disease, monitors treatment effects, and determines risk for clinical complications. These simplified tests are easier to administer and less invasive, thus having the potential to be more widely accepted by care providers administering the test and by patients receiving the test.

INTRODUCTION

It is estimated that over 100 million Americans have liver disease; however, only 4.5 million (1.8%) have been diagnosed. ¹ Left untreated, liver disease can lead to poor outcomes including liver cancer, decompensation, liver failure, and death. Currently, primary care providers, gastroenterologists, and hepatologists rely on biomarker or elastography surrogates, insensitive blood tests, or liver biopsy and invasive radiologic procedures to define a patient's liver health.

Current noninvasive liver tests measure fibrosis, inflammation, or steatosis rather than liver function. Elastography (liver stiffness) is linked to the degree of fibrosis; however, its reliability is operator‐dependent, and the relationship is complicated by hepatic steatosis, obesity, hepatic congestion, hepatic inflammation, and fasting versus fed status. Magnetic resonance imaging‐derived proton density fat fraction (MRI‐PDFF) can noninvasively measure steatosis, but the accuracy of MRI‐PDFF measurements in patients with advanced liver disease may be affected by the severity of fibrosis. ² , ³ Blood‐based biomarkers such as the ELF™ (Enhanced Liver Fibrosis) test ⁴ target hepatic fibrosis and do not measure liver function. Despite their colloquial label as “liver function tests”, ALT and AST indirectly measure hepatocellular injury or inflammation—not liver function. ⁵ The static measurements of bilirubin, albumin, and prothrombin time/INR detect hepatic dysfunction only in late stages of disease. Each of these markers can be influenced by non‐hepatic factors, and none of them measure hepatic function directly.

Clinical trials utilize liver histology and measurements of hepatic venous pressure gradient (HVPG) as end points. However, these methods are invasive, expensive, impractical for routine baseline testing or monitoring, and, furthermore, suffer from lack of reproducibility and sampling error. ⁶

For clinical use, invasive procedures such as liver biopsy or HVPG are not well accepted by patients and are compromised by the aforementioned technical challenges. Additionally, they are not reliable for use within an individual to assess liver disease severity, changes in liver health, or for predicting clinical outcomes. ⁷ There is a strong need for accurate and minimally invasive assessments of liver function to assist healthcare providers, patients, and payers in managing the full spectrum of liver diseases.

The HepQuant SHUNT test assesses both hepatocyte function and the portal circulation, making it a comprehensive diagnostic tool that addresses the pathogenesis of all types of liver disease. However, the SHUNT test (SHUNT V1.0) requires both oral and intravenous (IV) cholate doses and six peripheral venous blood samples taken over 90 min. Potential sources of error related to the IV injection and multiple timed blood samples include poor IV access (difficulty in locating vein or catheter dislodges from vein), extravasation, errors in the timing of sampling, and variability of skill in administering the test.

This work explores the reproducibility of HepQuant DuO, an oral‐only cholate challenge test, and other simplified versions of the prototypical SHUNT V1.0 test. The simplified tests, SHUNT V1.1, SHUNT V2.0, and DuO, reduce blood draws to four (SHUNT V1.1) or two (SHUNT V2.0, DuO) made possible by a compartmental model, ⁸ and in the case of DuO, eliminates the IV injection. We measured the within‐individual reproducibility of the simplified tests.

METHODS

This study is a retrospective analysis of data from two reproducibility studies of the HepQuant SHUNT test (236 tests completed in 94 subjects): the HepQuant Reproducibility (REPRO) study ⁹ and the Primary Sclerosing Cholangitis (PSC) study.

Subjects

The REPRO study comprised three groups of subjects: healthy persons without liver disease, patients with metabolic dysfunction‐associated steatohepatitis (MASH), and patients with hepatitis C virus (HCV) infection. The MASH group was diagnosed based on risk factors (obesity, diabetes, and metabolic syndrome), negative tests for other liver diseases, and fibrosis stage determined by either liver biopsy or transient elastography. The HCV group was diagnosed based on a positive HCV history through nucleic acid testing and METAVIR fibrosis stage determined by liver biopsy. Three separate HepQuant SHUNT tests were conducted on three different days within a 30‐day period.

Data obtained from the PSC study ¹⁰ were analyzed retrospectively. Reproducibility was assessed by conducting two baseline SHUNT tests in 46 subjects representing the clinical spectrum of PSC.

The REPRO and PSC studies were conducted according to The Code of Ethics of the World Medical Association (Declaration of Helsinki). Informed consent was obtained from all subjects. The REPRO study was approved by the Colorado Multi‐Institutional Review Board and registered at ClinicalTrials.gov (NCT01579162). The PSC study was approved by the Institutional Review Board of the University of Colorado Denver, and cholate isotopes were studied under IND 65121 (13C‐CA) and IND 65123 (d4‐CA).

HepQuant Test versions

The HepQuant Test versions are displayed in Table 1 and Figure 1 and are described in the subsequent sections.

TABLE 1.

Overview of analysis methods in terms of number and administration route of cholate isotope doses, number of blood samples, and duration of the testing period in minutes.

Test version	No. of isotope doses a	No. of samples	Duration of testing period
SHUNT V1.0	1 oral + 1 IV	6	90 min
SHUNT V1.1	1 oral + 1 IV	4	90 min
SHUNT V2.0	1 oral + 1 IV	2	60 min
DuO	1 oral	2	60 min
STAT	1 oral	1	60 min

^a In prior studies, the oral cholate isotope was d4‐CA, and the IV cholate isotope was 13C‐CA.

FIGURE 1.

Pharmacokinetic clearance curves for various versions of the HepQuant test on a MASH subject with F3 fibrosis. SHUNT V1.0 (a) measures the concentration of orally administered (PO) d4‐CA and intravenously (IV) administered 13C‐CA in systemic circulation across 90 min. SHUNT V1.1 (b) uses the patient's physical characteristics to calculate the initial 13C‐CA concentration, eliminating the need for sampling at 5 min. SHUNT V2.0 (c) calculates oral and IV clearances using samples collected at 20 and 60 min. DuO (d) calculates portal clearance from samples collected at 20 and 60 min and systemic clearance by derived IV.

SHUNT V1.0

SHUNT V1.0 ¹¹ (Figure 1a) describes IV clearance by noncompartmental exponential fits and oral clearance by a cubic spline fit to calculate the areas under the IV and oral curves (AUC_IV, AUC_Oral), respectively. Peripheral venous blood is obtained at 0, 5, 20, 45, 60, and 90 min within specified time windows, and concentrations of administered cholates are measured by liquid chromatography/mass spectrometry (LC/MS). After IV injection, residual 13C‐CA in the sampling catheter and sampling blood beyond the time window, especially for the 5‐min sample (±1 min), were sources of error in a minority of tests but may have compromised results in those tests.

SHUNT V1.1

SHUNT V1.1 (Figure 1b) eliminates the 5‐min sample from the calculations. The initial 13C‐CA concentration at 0 min is calculated from administered dose and estimation of blood volume from body mass index (BMI). ¹² The 5‐min 13C‐CA concentration is then estimated by log‐linear regression between the 0‐ and 20‐min concentrations, and the 5‐min oral concentration is approximated as 15% of the 20‐min d4‐CA concentration. The AUC_IV and AUC_Oral are then estimated by the same equations as in SHUNT V1.0.

SHUNT V2.0

SHUNT V2.0 (Figure 1c) implements (1) a compartmental model of portal cholate clearance that uses assumptions of liver flow and physiology to predict oral clearance curves and (2) noncompartmental exponential fits to systemic cholate clearance.

The compartmental model (Figure 2) for measuring portal clearance describes the flow between systemic (S), portal (P), and liver (L) compartments represented by a system of first‐order ordinary differential equations (Equations (1), (2), (3)), where q is the flow rate between compartments, V is the volume of the compartment, C is the concentration of d4‐CA in the compartment, Cl _H is the hepatic clearance, and D _PO,rate is the rate of orally administered d4‐CA entering the portal compartment.

(1)

$$\begin{array}{c}\frac{\mathrm{d}{C}_P}{\mathrm{d}t}=\frac{1}{V_P}\left(D_{\mathrm{PO},\text{rate}}+q_{\mathit{SP}}·C_S-q_{\mathit{PL}}·C_P-q_{\mathit{PS}}·C_P\right)\hfill \end{array}$$ (2)

$$\begin{array}{c}\frac{\mathrm{d}{C}_L}{\mathrm{d}t}=\frac{1}{V_L}\left(-{\mathrm{Cl}}_H·C_L-q_{\mathit{LS}}·C_L+q_{\mathit{PL}}·C_P+q_{\mathit{SL}}·C_S\right)\hfill \end{array}$$ (3)

FIGURE 2.

Compartmental model of dual cholate clearance describes the transfer of cholate between systemic, portal, and liver compartments. Cl_H, hepatic clearance; D _IV, intravenous dose of 13C cholate (SHUNT V2.0); D _PO, oral dose of d4‐CA (SHUNT V2.0 and DuO); q _SP, splanchnic arterial flow rate; q _PS, portal systemic shunt flow rate; q _PL, portal venous flow rate; q _LS, hepatic return to systemic circulation flow rate; q _SL, hepatic arterial flow rate (McRae et al. ⁸ )

The system of ordinary differential equations is solved numerically, and the concentration of d4‐CA in the systemic compartment is integrated over 180 min to calculate the AUC_Oral.

The noncompartmental analysis for measuring systemic clearance involves the exponential fits of systemic cholate clearance using the 20‐ and 60‐min 13C‐CA concentrations. The exponential fits split the curve into three clearance phases: fast (Y ₀), moderate (Y ₁), and slow (Y ₂). Equations (4), (5), (6) define the systemic concentration of 13C‐CA through time.

$$Y_0=C_0·e^{-k_{\text{fast}}·t}$$ (4)

$$Y_1=\frac{C_{20}}{e^{-k_{\mathrm{mod}}·T_{20}}}·e^{-k_{\mathrm{mod}}·t}$$ (5)

$$Y_2=\frac{C_{60}}{e^{-k_{\text{slow}}·T_{60}}}·e^{-k_{\text{slow}}·t}$$ (6)

Here, t is time (0–20, 20–60, and 60–180 min for Y ₀, Y ₁, and Y ₂, respectively), C₀ is the initial concentration of 13C‐CA, C₂₀ and C₆₀ are the measured 20‐ and 60‐min concentrations of 13C‐CA, and T₂₀ and T₆₀ are the actual collection times of the 20‐ and 60‐min blood sample. The systemic concentration of 13C‐CA is then integrated to calculate the AUC_IV. See the Supplemental Materials S1 for a more detailed description of both the compartmental analysis for portal clearance and the noncompartmental analysis for systemic clearance.

DuO

DuO (Figure 1d) is an oral‐only test involving administration of one oral dose (d4‐CA at 0 min) and collection of two blood samples (20 and 60 min) to quantify portal hepatic filtration rate (HFR) and estimate systemic HFR. These analyses involve the same compartmental model of portal cholate clearance as SHUNT V2.0. The systemic clearance is calculated by first estimating the derived IV concentrations at 20 and 60 min using linear regression models (Table S1) with body weight, BMI, the actual time of the 20‐min blood sample, and d4‐CA concentrations at 20 and 60 min (see Supplemental Materials S1 for a detailed description of the linear models and the regression coefficients). The derived IV concentrations are then used in the same noncompartmental analysis as SHUNT V2.0.

STAT

STAT 1.0 is simply the d4‐CA concentration at 60 min normalized to 75 kg body weight by the calculation ([d4‐CA] × (kg body weight/75 kg)). The value derived from the STAT test may be used by itself or in the estimation of portal HFR and DSI (Figure S1).

HepQuant Test parameters

The following parameters are calculated by the SHUNT versions and DuO and have previously demonstrated associations with liver function:

• Portal hepatic filtration rate (HFR_P) is the portal clearance adjusted for body weight (BW) (units mL min⁻¹ kg⁻¹) calculated by Equation 7, where D _PO is the oral dose.

$${\mathrm{HFR}}_P=\frac{D_{\mathrm{PO}}}{{\mathrm{AUC}}_{\text{Oral}}·\mathrm{BW}}$$ (7)

• Systemic hepatic filtration rate (HFR_S) is the measured systemic clearance adjusted for BW (units mL min⁻¹ kg⁻¹) calculated by Equation 8, where D _IV is the intravenous dose.

$${\mathrm{HFR}}_S=\frac{D_{\mathrm{IV}}}{{\mathrm{AUC}}_{\mathrm{IV}}·\mathrm{BW}}$$ (8)

• SHUNT% is the estimated absolute bioavailability of orally administered d4‐CA, or equivalently the ratio of systemic and portal HFRs (Equation 9).

$$\begin{array}{c}\text{SHUNT}\%=\frac{{\mathrm{AUC}}_{\text{Oral}}·D_{\mathrm{IV}}}{{\mathrm{AUC}}_{\mathrm{IV}}·D_{\mathrm{PO}}}=\frac{{\mathrm{HFR}}_{\mathrm{S}}}{{\mathrm{HFR}}_{\mathrm{P}}}\hfill \end{array}$$ (9)

• DSI is a score indicative of overall liver function comprising both portal and systemic HFRs. The DSI conveys unique information regarding fibrosis stage and clinical stages of cirrhosis, and modeling these outputs for prediction of clinical outcomes produced the DSI in Equation 10. ¹¹ , ¹³ , ¹⁴ , ¹⁵ Here, HFR_P,max and HFR_S,max are the upper limits of clearance for healthy controls, and A is a factor to scale DSI from 0 to 50.

$$\begin{array}{c}\mathrm{DSI}=A·\sqrt{{\left(\ln \left[\frac{{\mathrm{HFR}}_{\mathrm{P},\max }}{{\mathrm{HFR}}_{\mathrm{P}}}\right]\right)}^2+{\left(\ln \left[\frac{{\mathrm{HFR}}_{\mathrm{S},\max }}{{\mathrm{HFR}}_{\mathrm{S}}}\right]\right)}^2}\hfill \end{array}$$ (10)

• Hepatic Reserve is a numerical index representing overall hepatic health with values between 0 and 100 (Equation 11). Here, Hepatic Reserve uses HFR_P and HFR_S indexed to lean controls minus one standard deviation (HFR_P,lean and HFR_S,lean) with scaling factor, B, to scale Hepatic Reserve from 100 to 0. Any HFR_P value above HFR_P,lean is set to the HFR_P,lean value, and similarly any HFR_S value above HFR_S,lean is set to the HFR_S,lean value. Thus, any patients with both HFR values greater than these limits are calculated to have a Hepatic Reserve of 100.

$$\begin{array}{c}\text{Hepatic Reserve}\hfill \\ =100-B·\sqrt{{\left(\ln \left[\frac{{\mathrm{HFR}}_{\mathrm{P},\text{lean}}}{{\mathrm{HFR}}_{\mathrm{P}}}\right]\right)}^2+{\left(\ln \left[\frac{{\mathrm{HFR}}_{\mathrm{S},\text{lean}}}{{\mathrm{HFR}}_{\mathrm{S}}}\right]\right)}^2}\hfill \end{array}$$ (11)

• STAT may be derived from SHUNT, DuO, or STAT tests as the d4‐CA concentration at 60 min normalized to 75 kg body weight.

Reproducibility analysis

The reproducibility of test parameters was assessed through intraclass correlation coefficient (ICC), coefficient of variation (CV), and minimum detectable difference (MDD). The ICC method was single rater/measurement, two‐way mixed effects model for absolute agreement among measurements ¹⁶ , ¹⁷ with one‐sided test for the lower acceptable limit (ICC >0.7) defined by a reduction in accuracy of <5%. ⁹ Acceptable reproducibility was determined by ICC above the lower acceptable limit of 0.7 (p < 0.05). The MDD, interpreted as the smallest detectable difference in value of a test parameter before and after experimental manipulation (e.g., treatment), which is larger than would be expected by chance, ¹⁸ was calculated for each test parameter. ¹⁹ , ²⁰ The reproducibility was assessed for each study independently (three paired tests per subject for REPRO and two paired tests per subject for PSC) and for both studies combined. For the combined analysis, three pairings of two tests per subject were derived for the REPRO study data according to the study visits (V), that is, V1–V2, V1–V3, and V2–V3.

RESULTS

Subject characteristics

REPRO study

The REPRO study comprised 48 subjects who were categorized into three groups: controls (n = 16), MASH (n = 16), and HCV (n = 16). Subject characteristics (Table 2) and standard laboratory test results (Table S2) from the REPRO study ⁹ were reproduced here. The control group had an average age of 32.9 ± 12.0 years, a balanced male‐to‐female ratio of 8:8, and a normal BMI of 23.0 ± 2.2. The MASH group had an average age of 49.8 ± 11.6 years, had a male‐to‐female ratio of 7:9 (M:F), and a high BMI of 32.5 ± 5.9. The HCV group had an average age of 55.6 ± 6.7 years, a male‐to‐female ratio of 13:3, and an intermediate BMI of 28.2 ± 4.1. Most subjects (40 of 48) were non‐Hispanic White. Compared with controls, both the MASH and HCV groups had significantly higher levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma‐glutamyl transferase (GGT), and glucose, as well as lower platelet counts. The MASH and HCV groups were classified based on fibrosis stage as either F0–F2 (early) or F3–F4 (advanced), with eight subjects in each stage. All F4 subjects had well‐compensated liver disease without history of clinical complications.

TABLE 2.

Selected characteristics of study subjects from the HepQuant Reproducibility (REPRO) Study and Primary Sclerosing Cholangitis (PSC) study. Data are displayed as number of subjects or mean ± SD values.

Characteristics	REPRO study c			PSC study
	Controls	MASH	HCV
No. of subjects	16	16	16	47
Age (years)	32.9 ± 12.0	49.8 ± 11.6	55.6 ± 6.7	48.8 ± 12.9
Gender (M:F)	8:8	7:9	13:3	36:11
BMI (kg/m2)	23.0 ± 2.2	32.5 ± 5.9	28.2 ± 4.1	26.2 ± 3.9
Race/ethnicity (NHW:H:B:U)	12:2:1:1	14:2:0:0	14:1:1:0	43:1:3:0
Fibrosis stage, N subjects (F0:F1:F2:F3:F4) a	NA	0:4:4:5:3	3:2:3:4:4	NA
Fibrosis score a	NA	2.4 ± 1.1	2.3 ± 1.5	NA
MELD score b	NA	7.0 ± 1.1	7.9 ± 2.0	9.5 ± 4.2
CTP score b	NA	5.0 ± 0.0	5.4 ± 0.7	6.0 ± 1.5

Abbreviations: B, Black race; BMI, body mass index; CTP, Child–Turcotte–Pugh; H, Hispanic; HCV, cases with chronic hepatitis C; M:F, male‐to‐female ratio; MELD, model for end‐stage liver disease; MASH, cases with metabolic dysfunction‐associated steatohepatitis; NA, not applicable—healthy controls did not undergo liver biopsy and none had clinical disease, and fibrosis stage was not reported in the PSC study; NHW, non‐Hispanic White; U, undocumented.

^a METAVIR in HCV cases; BRUNT/KLEINER in MASH cases; 2 MASH cases had F3 fibrosis by transient elastography.

^b MELD and CTP scores were calculated for the F3/F4 subgroup only and in the PSC study.

^c From the REPRO Study (Burton et al.).

PSC study

Subject characteristics (Table 2) and standard laboratory test results (Table S2) from the PSC study were recorded. Of the 47 participants in the study, 43 were non‐Hispanic White, 3 were African American, and 1 was Hispanic. The male‐to‐female ratio was 36:11, with an average age of 48.8 ± 12.9 years, body weight of 81.3 ± 14.9 kg, and BMI of 26.2 ± 3.9 kg/m². Inflammatory bowel disease (IBD) was present in 70% (33) of the participants, with 76% (25) having ulcerative colitis. Among the participants, 36% (17) had a history of varices, ascites, or encephalopathy, with 15 having varices, 2 having variceal hemorrhage, 5 having diuretic‐responsive ascites, and 8 having treatment‐responsive encephalopathy. Radiologic reports indicated that 36% (17) had splenomegaly, and 28% (13) had a platelet count <140,000 μL⁻¹. Additionally, 19% (9) had a history of jaundice, 38% (18) had a history of bacterial cholangitis, and 72% (28 out of 39) of those with endoscopic retrograde cholangiopancreatography (ERCP) reports had undergone stricture dilation, stenting, or both. Treatment with ursodeoxycholic acid was administered to 66% (31) of the participants, while 40% (19) were taking IBD medications.

Reproducibility

REPRO study

In the REPRO study (Table 3) (n = 48 subjects), all test parameters were reproducible (ICC > 0.7, p < 0.05) for test versions SHUNT V1.1, SHUNT V2.0, and DuO. SHUNT% did not meet reproducibility criteria for SHUNT V1.0. STAT, Portal HFR, and DSI were reproducible for test version STAT. Across all test versions, the MDD ranged from 1.25 to 1.60 for DSI, 3.35% to 10.21% for SHUNT%, 5.18 to 6.05 for Portal HFR, and 1.71 to 2.65 for Hepatic Reserve.

TABLE 3.

Reproducibility of test parameters calculated by various test versions for all subjects in the HepQuant reproducibility study (n = 48).

Parameter	Test version	CV	MDD	ICC (95% CI)	p‐value a
DSI	SHUNT V1.0	11.07%	1.47	0.94 (0.90–0.96)	<0.001
	SHUNT V1.1	9.62%	1.25	0.94 (0.91–0.97)	<0.001
	SHUNT V2.0	10.36%	1.43	0.94 (0.90–0.96)	<0.001
	DuO	10.64%	1.60	0.93 (0.89–0.96)	<0.001
SHUNT% (%)	SHUNT V1.0	14.93%	10.21	0.74 (0.63–0.84)	0.2120
	SHUNT V1.1	11.92%	5.83	0.85 (0.78–0.91)	<0.001
	SHUNT V2.0	12.64%	6.22	0.84 (0.76–0.90)	0.0011
	DuO	9.18%	3.35	0.90 (0.84–0.94)	<0.001
Portal HFR (mL min−1 kg−1)	SHUNT V1.0	10.40%	5.18	0.84 (0.76–0.90)	<0.001
	SHUNT V1.1	10.48%	5.36	0.84 (0.76–0.90)	0.0013
	SHUNT V2.0	11.59%	6.05	0.82 (0.73–0.89)	0.0056
	DuO	11.59%	6.05	0.82 (0.73–0.89)	0.0056
Systemic HFR (mL min−1 kg−1)	SHUNT V1.0	11.15%	1.21	0.82 (0.73–0.89)	0.0068
	SHUNT V1.1	6.48%	0.47	0.90 (0.84–0.94)	<0.001
	SHUNT V2.0	6.36%	0.46	0.90 (0.85–0.94)	<0.001
	DuO	4.47%	0.25	0.94 (0.90–0.96)	<0.001
Hepatic Reserve (%)	SHUNT V1.0	2.16%	1.71	0.96 (0.94–0.98)	<0.001
	SHUNT V1.1	2.30%	1.92	0.96 (0.94–0.98)	<0.001
	SHUNT V2.0	2.54%	2.54	0.95 (0.92–0.97)	<0.001
	DuO	2.60%	2.65	0.95 (0.91–0.97)	<0.001
STAT (μM)	STAT	18.20%	0.19	0.90 (0.84–0.94)	<0.001
STAT‐estimated DSI	STAT	8.34%	2.1	0.88 (0.81–0.92)	<0.001
STAT‐estimated Portal HFR (mL min−1 kg−1)	STAT	13.53%	5.18	0.82 (0.73–0.89)	0.0066

Abbreviations: CV, coefficient of variation; DSI, disease severity index; HFR, hepatic filtration rate; ICC, intra‐class correlation coefficient with 95% confidence interval and p‐value testing ICC > 0.7; MDD, minimum detectable difference.

^a Bold values represent statistical significance (p < 0.05).

Reproducibility of test parameters was assessed across fibrosis stages (n = 16 subjects for each fibrosis group and each subject category). Test parameters were calculated from AUCs of the serum concentrations of cholate isotopes. AUCs increased with increasing stage of fibrosis and were lowest in healthy persons and in patients with low stages of fibrosis. Low AUCs were subject to greater variance. As a result, ICCs were higher for every test parameter and test version with increasing stage of fibrosis (Table S3) and in subjects with disease compared with controls (Table S4). Disease Severity Index and Hepatic Reserve met reproducibility criteria across F0‐F2 fibrosis for SHUNT V1.0 and SHUNT V1.1 and across F3‐F4 for all test versions (p < 0.001). Point estimates of ICCs generally exceeded the reproducibility cutoff of 0.70 across all test parameters and test versions although not by statistically significant (p < 0.05) amounts in every case.

PSC study

In the PSC study (Table 4) (n = 46 subjects), all test parameters, except Systemic HFR, met criteria for reproducibility (ICC > 0.7, p < 0.05) across all test versions (p < 0.001). Across all test versions, MDD ranged from 1.89 to 2.37 for DSI, 4.60% to 5.80% for SHUNT%, 2.65 to 3.76 for Portal HFR, and 4.27 to 5.08 for Hepatic Reserve.

TABLE 4.

Reproducibility of test parameters calculated by various test versions for all subjects in the Primary Sclerosing Cholangitis study (n = 46).

Parameter	Test version	CV	MDD	ICC (95% CI)	p‐value a
DSI	SHUNT V1.0	8.28%	2.03	0.93 (0.88–0.96)	<0.001
	SHUNT V1.1	8.13%	1.89	0.94 (0.89–0.96)	<0.001
	SHUNT V2.0	9.61%	2.37	0.92 (0.86–0.95)	<0.001
	DuO	9.24%	2.19	0.92 (0.87–0.96)	<0.001
SHUNT% (%)	SHUNT V1.0	10.69%	5.80	0.92 (0.86–0.96)	<0.001
	SHUNT V1.1	9.79%	4.64	0.93 (0.88–0.96)	<0.001
	SHUNT V2.0	10.76%	5.80	0.91 (0.85–0.95)	<0.001
	DuO	9.00%	4.60	0.92 (0.86–0.96)	<0.001
Portal HFR (mL min−1 kg−1)	SHUNT V1.0	11.30%	2.65	0.90 (0.83–0.94)	<0.001
	SHUNT V1.1	11.28%	2.67	0.90 (0.82–0.94)	<0.001
	SHUNT V2.0	13.23%	3.76	0.86 (0.77–0.92)	0.0017
	DuO	13.23%	3.76	0.86 (0.77–0.92)	0.0017
Systemic HFR (mL min−1 kg−1)	SHUNT V1.0	10.38%	1.40	0.79 (0.64–0.88)	0.0984
	SHUNT V1.1	8.49%	0.87	0.81 (0.69–0.89)	0.0341
	SHUNT V2.0	8.79%	0.96	0.79 (0.65–0.88)	0.0820
	DuO	5.86%	0.44	0.88 (0.79–0.93)	<0.001
Hepatic reserve (%)	SHUNT V1.0	4.70%	4.32	0.94 (0.90–0.97)	<0.001
	SHUNT V1.1	4.71%	4.27	0.94 (0.90–0.97)	<0.001
	SHUNT V2.0	5.08%	5.08	0.93 (0.88–0.96)	<0.001
	DuO	5.04%	4.98	0.93 (0.88–0.96)	<0.001
STAT (μM)	STAT	20.87%	0.33	0.92 (0.86–0.95)	<0.001
STAT‐estimated DSI	STAT	9.62%	3.10	0.89 (0.80–0.94)	<0.001
STAT‐estimated Portal HFR (mL min−1 kg−1)	STAT	16.91%	8.23	0.70 (0.52–0.82)	0.4989

Abbreviations: CV, coefficient of variation; DSI, disease severity index; HFR, hepatic filtration rate; ICC, intra‐class correlation coefficient with 95% confidence interval and p‐value testing ICC > 0.7; MDD, minimum detectable difference.

^a Bold values represent statistical significance (p<0.05).

Combined

In an analysis of the combined results from both the REPRO and PSC studies (Table S5) (n = 94 subjects, 236 tests), all test parameters met criteria for reproducibility (ICC > 0.7, p < 0.05) across all test versions (p < 0.001). Minimum detectable difference ranged from 1.39 to 1.72 for DSI, 3.6% to 7.8% for SHUNT%, 4.57 to 5.48 for Portal HFR, and 2.46 to 3.28 for Hepatic Reserve.

DISCUSSION

This study confirmed the reproducibility of HepQuant DuO, a noninvasive blood‐based test that only requires oral dosing of cholate and two blood samples at 20 and 60 min. Since DuO is embedded within the HepQuant SHUNT test, two formal reproducibility studies of SHUNT were used to quantify reproducibility and reliability of DuO. Establishing the within‐individual reliability of the DuO test is key to its clinical application.

The parameters of the SHUNT test have been independently linked to the risk of future clinical outcomes, such as in patients with various etiologies of cirrhosis, ²¹ ongoing HCV infection, ²¹ and PSC. ¹⁰ Further, the SHUNT test is associated with portal pressure ²² , ²³ and risk of toxicity and hepatic impairment in hepatocellular carcinoma treatment. ²⁴ The test has also been linked to improvement in liver function in several studies, including liver regeneration following living donor hepatectomy, ²⁵ gain of function following sustained viral response in advanced hepatitis C, ²⁶ , ²⁷ and treatment effect with obeticholic acid in MASH. ²⁸ Reproducibility studies have demonstrated DSI to be a reliable measure of liver function. ⁹ Recently, a pivotal study established a correlation between the results of the HepQuant test and the probability of developing large esophageal varices, ²⁹ , ³⁰ , ³¹ which are precursors to variceal hemorrhage and the deadliest complication of chronic liver disease.

One important finding of this study was that the reproducibility of the measurement of SHUNT% was improved by eliminating the 5‐min timepoint from the analysis. SHUNT V1.1 demonstrated better reproducibility in measuring SHUNT% than SHUNT V1.0. The initial clearance of IV‐administered 13C‐CA from systemic circulation exhibits a rapid exponential decrease in concentration. Thus, estimating the initial concentration based on data collected at 5 min is highly sensitive to variability from a variety of sources, including the timing of the dose administration and 5‐min blood sample collection, one‐ versus two‐arm catheter administration/collection, and residual cholate in the catheter line. In other words, small variations in the measurement and timing of the 5‐min 13C‐CA concentration may lead to significant changes in the AUC_IV and the HepQuant test parameters which depend on it (Systemic HFR, SHUNT%, DSI, Hepatic Reserve). Further, from recent experience, we estimated that over 50% of all test failures (i.e., non‐reportable results) from SHUNT V1.0 would be mitigated by eliminating the 5‐min sample. In this current study, we demonstrated that SHUNT V1.1 was more reproducible than SHUNT V1.0 across nearly every test parameter, with significant improvement in SHUNT% reliability. As a result, the MDDs for test parameters were lower with SHUNT V1.1 versus SHUNT V1.0. Improved test reliability offered by SHUNT V1.1 has the potential to detect smaller changes in therapeutic drug trials and/or require fewer subjects to achieve statistical power. For these reasons, we recognize SHUNT V1.1 as the reference method for comparing the next‐generation tests (SHUNT V2.0 and DuO).

A further simplification of the SHUNT test, SHUNT V2.0, was a reduction in the testing duration from 90 to 60 min and requirement for only two blood samples. Previously in the development of the analysis for SHUNT V1.0, a 2‐point version of the noncompartmental analysis was not accurate or reproducible. ¹¹ In contrast, SHUNT V2.0, using a compartmental model of portal clearance, which incorporates aspects of liver function and physiology, is accurate and reproducible (ICC > 0.7, p < 0.05). Relative to SHUNT V1.1, SHUNT V2.0 is simpler to administer and less invasive, thus having the potential to be more widely accepted by care providers administering the test and by patients receiving the test. The MDDs for test parameters were similar between SHUNT V1.1 and SHUNT V2.0. The SHUNT V2.0 test may be well‐suited for clinical scenarios where accurate measurements of systemic clearance are needed (e.g., monitoring portal‐systemic shunting or evaluation of new treatments targeting systemic clearance).

The ultimate goal of this study was to determine the reproducibility of test parameters generated from DuO, requiring only oral dosing of cholate and minimal sampling at 20 and 60 min. DuO mitigates several potential errors related to the IV injection (e.g., difficulty in locating vein for IV access, catheters dislodging from vein, extravasation) and may reduce variability associated with the skill in administering the test and collecting timed blood samples. Based on recent experience, we estimated that eliminating the IV injection would mitigate at least 80% of all failures (i.e., non‐reportable results) associated with the SHUNT V1.0 test. We found that DuO was highly reproducible with significance in ICC >0.7 for all test parameters. The MDDs for test parameters were similar between DuO and SHUNT V1.1. These results open the way for DuO to be used clinically where accurate measurements of portal clearance and general liver function testing are important, such as risk stratification, monitoring for treatment effects or disease progression, and predicting decompensation and adverse clinical outcomes. By using only oral dosing, DuO has the potential to broaden the utility of function testing in clinical practice.

Reproducibility was less consistent in MASH with low‐stage fibrosis, the group with greater portal HFR and lower AUC for orally administered cholate. This AUC has greater relative variance at low serum concentrations. One solution to this potential problem could be increasing the oral dose of administered cholate isotope to raise serum concentrations. Please note, however, that this observation is from a small subset analysis (n = 16) of the REPRO study. In a large analysis of approximately 1000 tests across a broad spectrum of fibrosis and disease severity, DSI from DuO was equivalent to DSI from SHUNT versions and met equivalency criteria by two one‐sided t‐tests and bioequivalence. ³² The potential source of error from low cholate concentrations is not a factor when assessing later stages of fibrosis or disease.

Variable gastric emptying and delayed absorption of the oral dose beyond 60 min could invalidate the oral clearance curves. This would manifest as very low to undetectable levels of cholate isotopes, which would be flagged by the laboratory during quality control and designated as an invalid result. Repeat testing would be required. Current experience suggests that this might occur in approximately 1% of tests.

We analyzed DuO and SHUNT simplified methods retrospectively from data that were collected using the SHUNT V1.0 method of test administrations. This study design is ideal since the DuO test and other simplified SHUNT versions are embedded within the SHUNT test, thus avoiding day‐to‐day variances when tests are performed on separate days or at different times. In addition, the LCMS assay has been extensively validated for ability to independently and accurately quantify individual cholate isotopes when both are co‐administered. In fact, it is likely that the more complicated tests (SHUNT V1.0 and SHUNT V1.1) are more susceptible to variability across sites, test administrators, and timing of blood samples, and thus, the reproducibility of the simplified tests (SHUNT V2.0 and DuO) may be underappreciated in the current study.

In conclusion, DuO and other simplified versions of the prototypical SHUNT V1.0 test represent advances in the noninvasive measurement of human liver function and physiology. DuO and SHUNT V2.0 are simple to administer liver function tests that will aid providers in defining severity of liver disease, tracking disease progression, monitoring treatment effects, and communicating this objective information to their patients.

AUTHOR CONTRIBUTIONS

M.P.M. and G.T.E. wrote the manuscript. All authors designed the research. M.P.M. and S.M.H. performed the research. M.P.M., S.M.H., and G.T.E. analyzed the data. M.P.M. and G.T.E. contributed new reagents/analytical tools.

FUNDING INFORMATION

The REPRO study was funded by Bioscience Discovery Evaluation Grant from State of Colorado, 10BGF‐13; and matching funds from Department of Medicine and Division of Gastroenterology and Hepatology, University of Colorado School of Medicine. The PSC cohort was studied under a research grant from the PSC Partners Foundation. The Agilent LCMS system used in this study was owned by HepQuant, LLC, and samples for this study were analyzed for cholate and cholate isotopes in the laboratory of G. T. Everson at the University of Colorado Denver via an Equipment Use Agreement between HepQuant, LLC and the University. The study sponsors had no role in study design, conduct, and interpretation of findings. The development of the next‐generation tests was funded by HepQuant LLC.

CONFLICT OF INTEREST STATEMENT

GTE (CEO) and SMH (CSO) are equity members and employees of HepQuant, LLC. They, in conjunction with the University of Colorado Denver, have issued and pending patents related to the dual cholate test (HepQuant SHUNT Test). MPM and JK are paid consultants for HepQuant LLC. MPM, SMH, and GTE have provisional patents related to administered cholate tests pending. All authors have read the journal's policy on conflicts of interest.

Supporting information

Data S1.

McRae MP, Kittelson J, Helmke SM, Everson GT. Within‐individual reproducibility of a dual sample oral cholate challenge test (DuO) and simplified versions of the HepQuant SHUNT test. Clin Transl Sci. 2024;17:e13786. doi: 10.1111/cts.13786