Liver Stiffness Increases Acutely During Sickle Cell Vaso-Occlusive Crisis

Abstract

Background & Aims

Acute vaso-occlusive crisis (VOC) in sickle cell disease (SCD) is an important cause of end-organ damage. It is estimated that 10–39% of VOC occurs with hepatic involvement. Current assessments of hepatic involvement during VOC are unsatisfactory. We investigated transient elastography (TE) as a marker of hepatic involvement, its relationship with histology, and biochemical markers during VOC.

Methods

SCD patients were evaluated with biochemical markers and TE at steady-state and during VOC. Change in TE and biochemical markers were correlated to length of hospital stay. When available, liver biopsy and tricuspid regurgitation velocity (TRV) at steady-state were correlated with TE.

Results

23 patients were evaluated (mean age=34.3 years, standard deviation=7.96). In 15 patients with liver biopsies, TE correlated with fibrosis (p=0.01) and TRV (p=0.0063), but not hepatic iron. Hemolysis biomarkers changed during VOC (p<0.022), but not alanine aminotransferase (ALT). Paired comparison of TE at steady-state and during VOC showed an increased from 6.2 to 12.3 kPa (p=0.0029). Increasing TE during VOC associated with increasing ALT and alkaline phosphatase (p=0.0088 and 0.0099, respectively). At steady-state, increasing inflammation on biopsy (p=0.0037) and TRV (p=0.0075) correlated with increasing TE during VOC. Increased hospital stay was associated with higher ALT (p=0.041), lower albumin (p=0.046), hemoglobin/hematocrit (p<0.0021) but not TE.

Conclusions

TE may identify patients with hepatic involvement during VOC independent of biochemical measures. Increase in TE may reflect both hepatic passive congestion and hepatic involvement during VOC. TE may serve as a physiological biomarker for hepatic features of VOC.

Introduction

Sickle cell disease (SCD) is one of the most common autosomal recessive disorders and is estimated to affect 1 out of every 70,000–100,000 individuals in the United States, 1 out of every 500 Black or African-American births and 1 out of every 36,000 Hispanic-American births.^1–3 SCD is a consequence of an abnormality at the sixth position of the β-globin chain of hemoglobin A, where a single nucleotide substitution occurs where valine replaces glutamic acid.^4,5 During periods of lowered oxygen saturation, this abnormality results in hemoglobin polymerization, leading to deformation of red blood cells, vaso-occlusion, ischemia/infarction and hemolysis.⁶ Although all organs can be affected, the most salient clinical features include chronic hemolytic anemia, periodic and painful vaso-occlusive crisis (VOC), infections due to asplenia, and death from potential multi-organ system failure.

As a consequence of SCD, various forms of acute and chronic hepatic injury have been described, ranging from hepatitis, hepatic iron overload, focal nodular hyperplasia, nodular regenerative hyperplasia, sickle cell hepatopathy, and acute hepatic infarction to cirrhosis.^7–14 Studies in patients that have been hospitalized due to acute vaso-occlusive crisis have estimated the frequency of liver involvement ranging from 10%-39% and an autopsy study of sickle cell patients has observed the occurrence of hepatic infarction in 34% of patients.^10,15–18 Clinical gastrointestinal manifestations include right upper quadrant pain, acute painful hepatomegaly, abdominal distention, fever, jaundice, and painful abdominal crisis.^17,19,20 Biochemically, serum transaminases abnormalities vary along with bilirubin levels.^10,19,21 Prior studies attempting to utilize biochemical parameters as tools to distinguish acute VOC with and without hepatic involvement have been unsuccessful, and specific correlations between clinical, biochemical and histological features of acute VOC remain unclear.^10,17,22 Additionally, the use of liver biopsy as a tool of hepatic assessment during acute VOC has proven to be a dangerous practice with resultant complications in up to 80% of patients including hemorrhage, organ failure and death.^14,23,24 Given the lack of clear definitions and unsatisfactory tools available for assessing hepatic involvement, the dangers of liver biopsy during acute VOC and the lack of a clear gold-standard, novel methods of delineating liver damage in VOC are needed. There are suggestions that such definitions and tools are clinically important. For example, ALT correlates with length of hospitalization for acute VOC.¹⁷

In order to further understand this, we prospectively investigated the utility of transient elastography (TE) as a marker of hepatic involvement in acute VOC in patients with sickle cell disease. TE has been evaluated in various liver diseases and has been most extensively validated in chronic hepatitis C patients.^25–28 Additionally, TE has been studied in patients with iron overload diseases such as β-thalassemia and primary hemochromatosis, however most studies have been on the identification of fibrosis in these cohorts.^29–31 Thus, no data exist on TE in patients with SCD at steady state or during acute VOC. Additionally, there is limited data on whether passive congestion due to increased right heart pressures plays a role in hepatic involvement during acute VOC. We attempted to determine the relationship between TE with liver histology, biochemical markers of liver injury, tricuspid regurgitation velocity (TRV) as a surrogate for pulmonary artery pressures, and length of hospital stay in a cohort of patients with SCD during sickle cell VOC.

Patients and Methods

Patients

All patients with SCD and acute VOC were admitted to the National Institutes of Health (NIH) Clinical Center for treatment of acute VOC between September 2009 and August 2012. During admission for the management of acute VOC, all patients prospectively underwent biochemical examination and transient elastography. Patients were also seen on an outpatient basis to obtain a steady state TE and transthoracic echocardiogram to obtain cardiac structure and flow measurements. Steady state was defined as at least 2 weeks away and a return to baseline laboratory parameters from an acute VOC in an asymptomatic patient. All patients provided written informed consent, and the study protocol and consent forms were approved by the institutional review boards of the National Institute of Diabetes and Digestive and Kidney Diseases and the National Heart, Lung and Blood Institute at the Clinical Center of the NIH and registered in ClinicalTrials.gov (#NCT0001971 and #NCT00081523). All coauthors had access to the study data and had reviewed and approved the final manuscript.

Laboratory Examination

All patients underwent a one-time thorough biochemical evaluation for viral hepatitis and other causes of liver disease. Those found to have other causes of liver disease were excluded. All patients underwent evaluation at steady state and during acute VOC with biochemical markers on the same day as TE. Markers include those that have been associated with acute VOC and sickle cell disease such as hemoglobin (HGB), hematocrit (HCT), ferritin, lactate dehydrogenase (LDH), and reticulocyte count. Liver associated laboratory tests such as ALT, AST, alkaline phosphatase (ALP), total bilirubin and direct bilirubin were also evaluated. Markers associated with hepatic synthetic function including albumin and prothrombin time (PT) were also measured.

Liver Biopsy Evaluation

In 15 patients, liver biopsies were previously obtained for clinically indicated purposes and were interpreted by a single hepatopathologist (DEK) who was blinded to the clinical data. All biopsies except one were obtained within 1 year of steady state examination (median: 44 days). Scoring of the liver biopsies utilized the modified histologic activity index (HAI) scale for inflammation and necrosis (range: 0–18) and the Ishak scoring system for fibrosis (range: 0–6).^32–34 Inflammatory scores included readings for periportal inflammation and necrosis (0–10), lobular inflammation and necrosis (0–4), and portal inflammation (0–4). Fibrosis ranged from 0 (none) to 1 to 2 (portal fibrosis only), 3 to 4 (bridging fibrosis), and 5 to 6 (incomplete and complete cirrhosis).

On light microscopy, two iron indices were also utilized to quantify hepatic iron overload. One scoring system, previously described by Rowe et al. (subsequently referred to as liver iron score), is scored from 0 to 4+ based on identification of stained iron via strengths of microscopic magnification with 4+ assigned to biopsies with dense aggregates of stainable iron identified under the lowest magnification and even with the naked eye.³⁵ The other scoring system, as described by Deugnier et al. (subsequently referred to as Deugnier score), scores iron deposits according to their size and cellular and lobular locations in the liver ascinus, leading to a summed score ranging from 0 to 60.³⁶

Transient Elastography Evaluation

All patients underwent transient elastography at steady state and during acute VOC within 48 hours of hospital admission. Transient elastography was performed via FibroScan© (Echosens, Paris) in accordance with technical recommendations.²⁸ Briefly, this is a rapid, painless, non-invasive, and reproducible bedside test that has been proposed for the assessment of hepatic stiffness.³⁷ The use of Fibroscan was approved by the United States Food and Drug Administration (FDA) on April 5, 2013 to aid clinical management in patients with liver disease.³⁸ Therefore, costs and availability of this procedure remain in flux with recent FDA approval. Results are expressed in kiloPascals (kPA) and correspond, according to the manufacturer’s recommendation to the mean of 10 valid measurements that can be completed in less than 5 minutes. Liver stiffness values range from 1.5 to 75 kPA with mean liver stiffness values in healthy volunteers 5.5 +/− 1.6 kPa.³⁹ A valid result was defined as at least 10 valid measurements, a success rate (the ratio of the total number of successful measurements to the total number of acquisitions) of at least 60% and interquartile range (reflecting the variability of validated measurements) <30% of the median liver stiffness measurement value. Results were considered unreliable if these criteria were not met. Failure of the procedure was defined as no valid measurement after at least 10 attempts.⁴⁰

Statistical Analysis

Continuous variables were reported as means (+/− standard deviation). Association of liver biopsy parameters with TE was evaluated by linear regression. Association of TRV at baseline with steady state TE was evaluated by Spearman’s correlation. Mean values of TE, ALT, AST, ALP, total bilirubin, direct bilirubin, albumin, PT, HGB, HCT, ferritin, LDH, and reticulocyte count at steady state and during acute VOC were compared via paired student’s t-test. Change in TE was defined as TE during VOC minus TE at steady state. Correlation between biochemical changes and histology with changes in TE after adjustment for steady state TE was compared by analysis of covariance. Correlation between TRV and change in TE from baseline to crisis was performed with Spearman’s correlation coefficient. Correlation between TE and biochemical parameters with hospital stay was performed with Pearson’s correlation. Significance was accepted at a P value of less than .05. Data analysis was performed using SAS, Version 9.1.3 (SAS Institute Inc., Cary, NC) software.

Results

Patient Demographics

Twenty-three patients with SCD admitted for acute VOC between September 2009 to August 2012 at the Clinical Center of the NIH were evaluated. These 23 patients included 7 men and 16 women, 10 African American, 7 African, and 6 Hispanic/Latino, with an average age of 34.3 years (standard deviation: 7.96 years) (Table 1). The median duration of hospital stay was 5.0 days. During the duration of the study, there was one death that occurred while the patient was at home. Although the exact circumstances leading to his death are not known, autopsy findings suggested that his death was due to an opioid overdose.

Table 1. Patient Demographics.

Clinical Features of 23 patients with Sickle Cell Disease and acute VOC

Feature	Number (%)
Males:	7 (30%)
Mean Age (standard deviation)	34.3 (7.96)
Number of patients with SS/SC/ST	22/1/0
Deaths	1 (4%)
Race
African American	10 (44%)
African	7 (30%)
Hispanic/Latino	6 (26%)
Number of Patients with Liver Biopsy	15 (65%)
Ishak Fibrosis Score 0–2	14 (93%)
Ishak Fibrosis Score 3–4	1 (7%)
Ishak Fibrosis Score 5–6	0 (0%)
Liver Biopsy Features	Median (Range)
HAI Total Inflammation (0–18)	3.0 (0–8)
Ishak Fibrosis (0–6)	0 (0–3)
Iron Score (0–4)	3.0 (1–4)
Liver Iron Concentration (normal: 200–2400 mcg/g dry weight)	6943 (1907–36416)
Deugnier Score	28 (10–46)

Abbreviations: VOC, vaso-occlusive crisis; HAI, Histologic Activity Index

Liver Biopsy Demographics and Histologic and Transient Elastography Associations

Of the 23 patients studied, liver biopsies were obtained on 15 patients at steady state (median time between biopsy and TE: 44 days). Results from these biopsies demonstrated a median HAI total inflammatory score of 3.0 (range 0–8) along with bridging fibrosis (Ishak fibrosis score 3–4) in 1, and no fibrosis or portal fibrosis (Ishak 0–2) in 14 (Table 1). No biopsy showed cirrhosis (Ishak fibrosis 5–6). Histologic indices of iron overload demonstrated hepatic iron overload with a median hepatic iron score of 3 (range 0–4), liver iron concentration of 6943 mcg/g dry weight (normal range 200–2400 mcg/g dry weight), and a Deugnier score of 28.0 (range 0–60).

A comparison of liver biopsy parameters with transient elastography was then performed (Supplemental Table 1). All patients met criteria for successful transient elastography measurements. Ishak fibrosis correlated with hepatic stiffness by transient elastography (p=0.01), however HAI inflammation did not. No indices of hepatic iron (histologic iron score, liver iron concentration and Deugnier score) correlated with transient elastography.

Comparison of Steady State Parameters and Vaso-occlusive Parameters

Twenty-three patients had transient elastography and biochemical parameters at steady state and during acute VOC compared (Figure 1 and Table 2). Mean transient elastography scores increased significantly during acute VOC (6.2 to 12.3 kPa, p=0.003). In evaluating the liver associated enzymes, ALT did not increase significantly, however changes in AST were significant (55 to 76 U/L, p=0.01). Alkaline phosphatase and direct bilirubin did not increase, however, total bilirubin increased significantly (2.2 to 3.9 mg/dL, p=0.001). With markers associated with hepatic synthetic function, albumin worsened significantly during acute VOC (3.9 to 3.7 g/dL, p=0.001), however PT was unchanged.

Figure 1.

Transient elastography (TE) measurements in kiloPascals (kPa) of 23 patients at steady state and during acute vaso-occlusive crisis. (A) Magnification from 0 to 15 kPa. (B) TE measurements of entire cohort shown in upper right hand corner.

Table 2.

Comparison of biochemical values at steady state and during vaso-cclusive crisis.

Feature	Steady State Mean Value	Crisis Mean Value	P Value
Transient Elastography (kPa)	6.2±4.8	12.3±11.9	0.003
ALT (U/L)	43±36	47±30	0.30*
AST (U/L)	55±50	76±48	0.01*
Alkaline Phosphatase (U/L)	107±51	120±59	0.16
Total Bilirubin (mg/dL)	2.2±1.4	3.9±5.3	0.001*
Direct Bilirubin (mg/dL)	0.47±0.24	1.49±3.33	0.24
Albumin (g/dL)	3.9±0.4	3.7±0.4	0.001
Prothrombin Time (sec)	15.9±1.8	17.0±4.4	0.17
Hemoglobin (g/dL)	8.9±1.9	8.0±1.5	0.002
Hematocrit (%)	25.5±5.8	22.7±4.1	0.0006
Ferritin (mcg/L)	2037±2218	3080±3449	0.59
Lactate Dehydrogenase (U/L)	451±290	650±458	0.002*
Reticulocyte (%)	8.5±5.3	11.2±6.1	0.02

Analysis by paired Students t-test

^*Log transformed prior to paired t-test

Markers of sickle cell anemia and acute VOC were also compared. As expected, hemoglobin (8.9 to 8.0 g/dL, p=0.002) and hematocrit (25.5 to 22.7 %, p=0.0006) significantly decreased whereas reticulocyte count (8.5 to 11.2 %, p=0.02) significantly increased during acute VOC, as did LDH (451 to 650 U/L, p=0.002).

Evaluation of individuals with hepatic involvement during acute VOC with TE, histology and biochemical changes

In patients who had increasing transient elastography values during acute VOC, biochemical parameters were evaluated to assess correlation with increasing liver stiffness (Table 3). ALT and alkaline phosphatase significantly increased (p=0.009 and 0.01, respectively). However, other hepatobiliary measures such as total and direct bilirubin and indices of hepatic synthetic function (albumin and PT) did not correlate with TE. Markers of sickling and VOC did not significantly change in this group with increasing liver stiffness. Additionally, neither the presence nor number of transfusions during acute VOC correlated with TE (data not shown). On comparison with steady state liver biopsy parameters, only HAI inflammation correlated with change in transient elastography (p=0.004) but not fibrosis or the various iron indices.

Table 3.

Association between biochemical changes and histology with changes in transient elastography after adjustment for steady state transient elastography.

Feature	Parameter Estimate	P Value
ALT (U/L)	2.92	0.009
AST (U/L)	3.97	0.0564
Alkaline Phosphatase (U/L)	2.99	0.01
Total Bilirubin (mg/dL)	0.06	0.42
Direct Bilirubin (mg/dL)	0.09	0.46
Albumin (g/dL)	−0.01	0.52
Prothrombin Time (sec)	0.38	0.36
Hemoglobin (g/dL)	0.08	0.10
Hematocrit (%)	0.22	0.10
Ferritin (mcg/L)	126.5	0.34
Lactate Dehydrogenase (U/L)	14.23	0.21
Reticulocyte (%)	0.38	0.06
HAI Inflammatory Score	0.95	0.004
Ishak Fibrosis Score	−0.10	0.26
Iron Score	0.21	0.95
Liver Iron Concentration (mcg/kg)	722.9	0.83
Deugnier Score	1.47	0.73

Analysis by ANCOVA regression, TE is the independent predictor and steady state TE is the covariate

Evaluation of TE and echocardiographic parameters of right heart ysfunction

Increases in TE and liver stiffness have been described with congestion due to intrahepatic portal hypertension.⁴¹ It has also been described (in animals and humans) to be elevated from extrinsic congestion such as heart failure.^42–44 In SCD, right ventricular pressures are known increase in VOC which may promote hepatic congestion.^45,46 Thus, although not previously explored in SCD, correlations were sought between TE and echocardiographic markers of right heart dysfunction.

All 23 patients underwent steady state echocardiography with a median time interval between echocardiography and steady state TE of 56 days. On echocardiography comparison with TE, and TRV and right ventricular systolic pressure (RVSP) were significantly correlated (p=0.0048 and 0.0066, respectively)(Supplemental Table 2). On comparison between steady state echocardiographic measurements with change in TE from steady state to crisis, TRV, right atrial area, RVSP and right atrial pressure significantly correlated with increasing TE (p=0.01, 0.048, 0.0003, and 0.004, respectively).

Evaluation of TE and biochemical parameters with length of hospital stay

Finally, the length of hospital stay was compared with the change in TE and biochemical parameters (Supplemental Table 3). Interestingly, although TE did not correlate with the length of hospital stay, an increasing ALT was associated with a longer duration of hospital stay (p=0.04). None of the other liver associated enzymes or indices of hepatic synthetic function correlated with hospital stay. However, a decrease in HGB and HCT was significantly associated with an increased duration of hospital stay (p=0.001 and 0.002, respectively) as was an increase in ferritin (p=0.0007).

Discussion

Patients with sickle cell disease are at risk for significant hepatic complications, and acute VOC with hepatic involvement is one with potential catastrophic consequences resulting in acute hepatic failure and a high mortality rate.^9,17,47,48 It has been suggested that this is due to widespread sickling with resultant obstruction in the sinusoids leading to hepatocyte anoxia and ischemia, and ultimately necrosis and death. Available definitions of hepatic involvement in acute VOC are imprecise given the overlap between sickle cell and hepatic serum biochemical parameters and the poor correlation of the serum biomarkers with acute VOC, which was corroborated in this study.^10,22 The traditional approach to define liver involvement in VOC has been with ALT elevation, however in liver disease, ALT is considered an imperfect tool. Significant hepatic damage can be missed if ALT is used as a sole marker. Ultimately, better definitions and markers could be utilized to understand the pathophysiology of hepatic involvement in acute VOC and its clinical implications, as well as the long-term sequelae of repeated hepatic involvement.

Given the lack of predictors of hepatic involvement during acute VOC, we evaluated the utility of transient elastography in identifying individuals with hepatic involvement. A relative strength of this manuscript is the availability of liver histology during steady state on a large percentage of patients in this cohort. Although a comparison of histology during acute VOC would have provided additional information, prior studies have shown the increased adverse outcomes of liver biopsy in acute VOC, therefore this analysis was not feasible.^14,23,24 However, histologic comparison with TE in this cohort remained consistent with previously described studies in patients with hepatitis C infection where hepatic fibrosis correlated with TE measurements.^25–28

On evaluation of histologic markers of iron overload including liver iron concentration, there was no significant correlation with TE. This has not been previously investigated in patients with sickle cell disease, however similar findings of poor correlation of hepatic iron overload and TE have been described in patients with hemochromatosis (HFE-linked iron overload) and thalassemia patients.^29–31 This suggests that the amount of iron in the liver does not influence liver stiffness as detectable by TE.

More interestingly, in comparison of steady state TE with measurements taken during acute VOC, hepatic stiffness was significantly increased during acute VOC. On further evaluation, the severity of liver stiffness also correlated with worsening markers of hepatic injury (ALT and ALP). Taken together, these results suggest that not only can TE identify patients with fibrosis in patients with SCD at steady state, but it is also able to identify individuals with hepatic involvement during acute VOC, perhaps also providing information regarding severity of hepatic involvement, although this remains to be proven. A potential explanation for the increase in liver stiffness, as detected by TE, may be due a temporary intrahepatic sinusoidal obstruction leading to increased liver stiffness.¹⁶

Furthermore it is noteworthy that steady state ALT, ALP and HAI inflammation on liver biopsy correlated with increase in TE during acute VOC. ALT and HAI inflammation on biopsy correlated (data not shown) suggesting that steady state necro-inflammation as measured by either ALT or HAI correlated with increase in TE during VOC. This implies that pre-existing hepatic inflammation may predispose patients to intra-hepatic vaso-occlusion during VOC.^49,50 It is the intra-hepatic vaso-occlusion that is likely to cause the increase in TE and the hepatic damage seen in a subset of patients.

In this study, TE shows a relationship to estimates of pulmonary artery pressure. Both baseline TE and change in TE from steady state to VOC correlated with baseline TRV and markers of right heart dysfunction. TRV is known to rise in VOC and is an accepted marker of elevated right ventricular pressure, which can promote hepatic congestion.^45,46 These correlations suggest that hepatic vascular congestion may also contribute to overall TE score. In prior animal studies, central venous pressure was described to directly influence liver stiffness.⁴³ These liver stiffness findings were later corroborated in 10 patients with decompensated congestive heart failure.⁴⁴ Thus, we corroborate these findings in sickle cell patients and demonstrate that patients with higher right-sided cardiac pressures are predisposed to increased TE measurements during VOC. Potential explanations for this finding could be that elevated right-sided pressures predispose to hepatic VOC, or that the increases in TE reflect further elevation in right-sided cardiac pressures during acute VOC, or a combination of the two.

Liver involvement has previously been reported and associated with longer hospitalization.¹⁷ In the cohort of patients in this manuscript, both ALT and ALP correlated with increasing TE during VOC. However, TE was also noted to increase in some who did not have substantial increases in ALT or ALP suggesting that increase in liver stiffness is not always reflected in biochemical changes. This is consistent with variable elevations of liver enzymes in other liver diseases independent of the degree of hepatic damage or impairment.

In evaluating the duration of hospital stay with TE and biochemical markers, it was interesting that a greater increase in ALT was associated with an increased duration of hospitalization. The correlation of elevated ALT and length of hospitalization suggests that those with hepatic involvement during acute VOC may be sicker than those without hepatic involvement, as also evidenced by increasing anemia. Unexpectedly, increasing liver stiffness was not associated with length of hospitalization, although increase in TE was correlated with increase in ALT. This lack of correlation may be due to the low number of patients evaluated in our study or a true lack of correlation. Further studies with larger numbers of SCD patients should be performed to evaluate this given the potential important clinical implications.

A shortcoming of this study is the low number of individuals evaluated and the fact that there were eight patients without steady state liver biopsies. The cross sectional nature of the study may be adequate to demonstrate that TE has utility in SCD however is not adequate for complete characterization of this modality. It is possible that the true ”steady state” if obtained before the crisis might have been lower than a “steady state” obtained after the crisis because of damage occurring during the crisis itself. However the potential implications remain true and the findings seem worthy of further exploration.

In summary, transient elastography has the potential to identify patients with hepatic involvement early on during acute VOC where biochemical measures have failed and those with the potential to have significant adverse outcomes. TE also has the potential to provide quantitative information on the severity of hepatic involvement in VOC and may be an important tool to study the biology of acute VOC.

Abbreviations

SCD

sickle cell disease

VOC

vaso-occlusive crisis

ALT

alanine aminotransferase

AST

aspartate aminotransferase

TE

transient elastography

NIH

National Institutes of Health

HGB

hemoglobin

HCT

hematocrit

LDH

lactate dehydrogenase

ALP

alkaline phosphatase

PT

prothrombin time

HAI

histologic activity index

kPa

kiloPascals