LiMAx system for assessing the functional capacity of the liver

Stockmann et al. (2018)

Study size, design and location

Phase III, multicentre, 2-arm, parallel-group, open-label RCT involving 148 randomised adult patients (≥18 years) with intrahepatic tumours scheduled for open liver resection of at least 1 segment. Done in 6 German academic centres specialised in complex liver surgery.

Intervention and comparator(s)

Intervention: LiMAx (n=58).

Comparator: Standard of care (n=60).

Key outcomes

In the intervention group, the LiMAx test was done before and after surgery for individual surgical planning and to prospectively determine the level of post-operative care, respectively. 62% (n=36/58) of patients in the LiMAx group were transferred directly to a general ward after surgery versus 2% (n=1/60) of patients in the control group (p<0.001). The rate of severe complications (grade ≥IIIa) was significantly lower in the LiMAx group compared with the control group (14% versus 28%; p=0.022). No statistically significant differences were seen for grade I or II complications. Compared with the control group, patients in the LiMAx group had a significantly shorter length hospital stay after surgery (10.6 versus 13.3 days; p=0.012), as well as shorter length of immediate care/ICU (0.8 versus 3.0 days; p<0.001).

Strengths and limitations

Study compared the technology to standard of care and was sufficiently powered to detect between-group differences. Randomisation helped reduce the risk of selection bias. Surgical techniques used were not significantly different between treatment groups.

The study was funded by the company and carried out in Germany. Centres specialised in complex liver surgery, and people with complex liver resections and those with previous resections or pre-existing fibrosis or cirrhosis were excluded from this study; results may not fully reflect real-world clinical practice. All outcomes were hospital process measures.

Jara et al. (2015a)

Study size, design and location

Retrospective analysis involving 1,170 consecutive patients having elective liver surgery between January 2006 and December 2011, a period spanning the introduction of the LiMAx algorithm in 2008 and 2009. By 2010, LiMAx and its algorithm were fully integrated into clinical practice at the study centre.

Intervention and comparator(s)

Intervention: LiMAx test.

No comparator.

Key outcomes

The proportion of complex liver surgeries increased from 29.1% in 2006 to 37.7% in 2011 (p=0.034). The proportion of patients with cirrhosis who were selected for liver surgery increased from 6.9% in 2006 to 11.3% in 2011 (p=0.039). Rates of liver failure after liver surgery decreased from 24.7% in 2006 to 0.9% in 2011 (p=0.014). Similar results were seen in an analysis for a propensity-score matched cohort, where reductions in the rates of liver failure (24.7% [n=77] versus 11.2% [n=35]; p<0.001) and related mortality (3.8% [n=12] versus 1.0% [n=3]; p=0.035) after liver surgery were shown.

Strengths and limitations

Data were from a large number of consecutive and non-selected patients submitted for partial liver surgery, overcoming potential selection bias. Liver surgeries followed a common surgical approach and most (70.8%) were done by 3 experienced liver surgeons, reducing the risk of performance bias.

The study was a single-centre analysis done in Germany. The potential effect of improved surgical techniques, anaesthetic care and intensive care nursing over the study period cannot be excluded. Data on the number of patients who were denied surgery based on actual LiMAx values were not available so this study only provides low-level evidence on diagnostic accuracy.

Jara et al. (2015b)

Study size, design and location

Single-centre prospective analysis of 167 patients with chronic liver failure without HCC evaluated for liver transplant between July 2009 and April 2013.

Intervention and comparator(s)

Intervention: LiMAx test.

MELD and ICG-PDR were evaluated as reference standards.

Key outcomes

Within 6 months of follow-up, 36 had liver transplant and 18 patients died. Median LiMAx values were significantly lower in liver transplant candidates who died versus those who survived (99 versus 50 micrograms per kilogram per hour), while ICG-PDR did not differ between the 2 patient groups (4.4 versus 3.5%/min; p=0.159). When identifiable cut-off values for predicting the probability of death within 6 months were applied, LiMAx had a higher negative predictive value (0.93), compared with ICG-PDR (0.90) and MELD (0.91).

Strengths and limitations

Study involved consecutive enrolment of all patients fulfilling inclusion criteria. Study had a follow-up of 6 months and included other liver function parameters as reference.

Single-centre study conducted in Germany. MELD scores were relatively low for liver transplant candidates, perhaps because of the exclusion of patients with acute onset of liver failure. Patients with HCC were excluded from the study. It is unclear whether all patients in this study were eligible for transplant and how the availability of suitable donors affected this.

Stockmann et al. (2010)

Study size, design and location

Observational trial involving 329 patients with liver tumours evaluated for liver surgery.

Intervention and comparator(s)

Intervention: LiMAx test.

Key outcomes

Blinded preoperative mean LiMAx values (study group) were significantly higher before resection (351 micrograms per kilogram per hour, n=139) versus before refusal (299 micrograms per kilogram per hour, n=29; p=0.009). In-hospital mortality rates were 38.1% (8/21 patients), 10.5% (2/19 patients) and 1.0% (1/99 patients) for post-operative LiMAx of <80 micrograms per kilogram per hour, 80 to 100 micrograms per kilogram per hour and >100 micrograms per kilogram per hour, respectively (p<0.0001). LiMAx levels <80 micrograms per kilogram per hour were associated with longer hospital stays and duration of intensive care. After developing a decision tree, its prospective preoperative application (routine group) also revealed higher LiMAx values before resection versus before refusal (257 versus 356 micrograms per kilogram per hour; p<0.0001). Intra-hospital mortality after surgery reduced from 9.4% in the blinded study group to 3.4% in the routine group (p=0.019).

Strengths and limitations

Selection of patients was not influenced by individual characteristics and medical personnel were blinded to preoperative LiMAx readouts.

Single-centre study done in Germany. Outcome and survival in the study group were only followed up until discharge from the hospital. There were significant between-group differences in aetiology and surgical procedures. Potential parameters that might bias the individual test result including obesity, tumour stage and the general condition of the patient cannot be excluded.

Lock et al. (2010)

Study size, design and location

Prospective, observational, pilot study involving 99 patients having deceased donor liver transplant between August 2005 and May 2007.

Intervention and comparator(s)

Interventions: the LiMAx test.

Comparators: ICG-PDR, and conventional biochemical parameters for the diagnosis of initial graft dysfunction.

Key outcomes

Patients with initial graft dysfunction had lower LiMAx values immediately after transplant (43 versus 184 micrograms per kilogram per hour; p<0.001) whereas ICG-PDR was only slightly decreased (11.8 versus 15.5 %/min; p=0.200). Significant differences were also seen for serum bilirubin, ammonia, glutamate dehydrogenase, and the INR (p<0.05). Multivariate analysis showed LiMAx to be the only single independent predictor of initial graft dysfunction (p<0.008). ROC analysis for LiMAx showed an AUROC curve of 0.960 (p<0.001). LiMAx was shown to detect initial graft dysfunction with a sensitivity of 1.0, specificity of 0.92, a positive predictive value of 0.53, and a negative predictive value of 1.0. The LiMAx and AST were used to detect primary non-function (PNF; n=3) on the first post-operative day. AUROC values were 0.992 (p=0.004) for LiMAx and 0.967 (p=0.006) for AST. Based on a combination of test results obtained immediately after transplant and on the first day, LiMAx was shown to detect diagnose PNF with a sensitivity of 1.0 and a positive predictive value of 1.0, while AST showed a sensitivity of 0.67 and a positive predictive value of 0.29. LiMAx showed significantly better diagnostic accuracy versus AST (p=0.031) for the diagnosis of PNF within 24 hours after transplant.

Strengths and limitations

Study compared LiMAx to ICG-PDR test and conventional biochemical parameters and post-operative outcomes were documented for 90 days. LiMAx and ICG-PDR were done by doctors who were not involved in clinical management and surgical re-intervention was decided independent of study results.

Data comes from 1 German centre. Immediate graft dysfunction was retrospectively defined from the patient's history. LiMAx cut-off values were determined post hoc from the same data set. Extrahepatic factors affecting LiMAx values such as co-administration of catecholamines and weight changes after transplant cannot be excluded.

Stockmann et al. (2009)

Study size, design and location

Single-centre, prospective observational study involving 64 adult patients (18 to 75 years) having liver surgery between August 2004 and February 2007.

Intervention and comparator(s)

Intervention: LiMAx test.

Key outcomes

Residual LiMAx values on post-operative day 1 showed significant correlation with residual liver volume (r=0.94; p<0.001). Multivariate analysis showed LiMAx to be the only predictor of liver failure (p=0.003) and mortality (p=0.004) on post-operative day 1. ROC analysis showed an AUROC of 0.99 for the prediction of both liver failure and liver failure-related death by LiMAx and an AUROC of 0.69 for severe complications. LiMAx was shown to predict liver failure related death with a sensitivity of 1.0 and a specificity of 0.93. An accurate calculation of the remnant liver function capacity before surgery was measured by combining computed tomography volumetry and LiMAx (r=0.85; p<0.001). No adverse events of the intravenous ¹³C-methacetin administration were observed during injection or follow-up.

Strengths and limitations

Patients were followed up for a total of 6 months and none of the patients withdrew consent. LiMAx measurements were validated in healthy volunteers and also during anhepatic phase of liver transplant. Histopathology evaluation was done by a blinded pathologist.

Data come from a single German centre.

Abbreviations: AST, aspartate aminotransferase; AUROC, area under the receiver operating characteristic; ICG-PDR, indocyanine green plasma disappearance rate; ICU, intensive care unit; INR, internal normalised ratio; HCC, hepatocellular carcinoma; MELD, Model for End-Stage Liver Disease; RCT, randomised controlled trial; ROC, receiver operating characteristic.