An investigation of liver blood flow in systemic inflammation.
MD thesis, University of Glasgow.
Full text available as:
Inflammatory stimuli such as infection or tissue injury will produce a local inflammatory response which may, if the inflammatory response is sufficiently large, spill over to produce a systemic inflammatory response. This is clinically characterized by a response in heart and respiratory rate, a temperature rise or fall and a white cell response. Where the systemic inflammatory response syndrome is due to proven infection this is classified as sepsis. If the inflammatory stimulus is removed or dealt with by the body or with medical treatment, the systemic inflammatory response may resolve with a return to homeostasis. In some patients the response does not resolve and they may progress to an anti-inflammatory state which predisposes to infection and poor wound healing or progress to multi-organ dysfunction syndrome, often resulting in cardiac, respiratory or renal failure. There is no specific treatment for multi-organ dysfunction syndrome and the practice of intensive care medicine has developed to support organ function in this period while the source of the illness is treated and the patient allowed time to recover. Mortality remains high in intensive care medicine with in hospital mortality around 30-40%.
Clinically markers of systemic inflammation are used to assess improvement or deterioration in condition. White cell count will be elevated in infection and is not a good indicator of systemic inflammation. The pro-inflammatory cytokines such as IL-6 mediate inflammation. These are secreted by activated macrophages of which 80% are resident in the liver as Kupffer cells. Cytokines act on the liver hepatocytes causing them to produce proteins. Plasma proteins that change in concentration with inflammation are named acute-phase proteins and the most clinically important of these is C-Reactive Protein. It is usually found in the plasma at concentrations of 3mg/l or less and values of less than 10mg/l are regarded as clinically unimportant. C-reactive protein is stable in plasma, does not increase with age, has little or no diurnal variation, changes rapidly with disease and has a wide range of abnormal values.
The liver is a source of cytokines, a target for cytokines, manufactures proteins and is the site of gluconeogenesis in systemic inflammation and thus plays a pivotal role in the process. In humans liver failure is rare in multi-organ dysfunction syndrome; however liver dysfunction is associated with poorer outcome in critical illness. As the liver is hypermetabolic the delivery of blood would appear to be important in critical illness.
Liver blood flow has been studied in systemic inflammation but the relative inaccessibility of the portal vein and the hepatic artery has made this a difficult task. The few studies that have been performed in critically ill patients have used the Fick principal to allow estimation of total liver blood flow by clearance of a marker in the blood by the liver. A requirement of this technique is hepatic venous catheterisation which is invasive and not without complications. Advances in ultrasound technology have allowed colour Doppler/ duplex scanning of the hepatic artery and portal vein trans-abdominally and allow non-invasive measurement of not only total liver blood flow but also its individual components. This technique has been shown to be reliable and reproducible and the results reported in Chapter 2 confirm with this in the authors’ hands.
The consensus from previous work is that total liver blood flow is increased in systemic inflammation. As mentioned above the relative contributions of the hepatic artery and portal vein in these changes has not been studied. In Chapter 3 a cross sectional study was performed in three different groups; controls, non-small cell lung cancers and acute alcoholic hepatitis. These three groups were shown to be different in levels of systemic inflammation with the controls least and the hepatitics most inflamed. Total liver blood flow was not significantly altered in these inflammatory disease states, nor was portal venous blood flow. In contrast, there was a significant increase in hepatic arterial blood flow which was related to increased systemic inflammation. The mediators of the increased hepatic arterial flow was not clear but it could be a hormonally mediated response, a cytokine mediated response or simply an intrinsic response to the increased metabolic demand that on the liver by systemic inflammation.
Previous liver blood flow studies in acute inflammation have been performed when inflammation is established and therefore it is difficult to define the chronological changes in liver blood flow. A previous study utilised ultrasound to measure the liver blood flow at 5 and 24 hours after onset of inflammation. This study reported an increase in hepatic arterial and portal venous blood flow at 5 hours compared to controls but no significant changes at 24 hours. Studies with infusions of interleukin-6 demonstrated an increase in liver blood flow that peaked at four hours in healthy volunteers. We therefore hypothesised that there were changes in hepatic arterial and portal venous blood flow within the first 6 hours of an inflammatory stimulus. In Chapter 4 serial hepatic arterial and portal venous blood flow measurements were made following the surgical trauma of lower limb arthroplasty. An immediate fall in portal venous flow was seen which was followed by an increase in hepatic arterial blood flow over the next four hours. By 24 hours the hepatic arterial blood flow was not significantly different to the pre-operative value and the portal venous flow was returning to normal, although remained statistically lower than pre-operatively.
The immediate fall in portal venous flow is likely to be hormonally mediated, as we know the adrenal medullary hormones that are released in response to tissue injury or trauma, will direct blood away from the gut to selectively perfuse the brain, heart, lungs, kidneys and muscle. The immediate effect would also favour a hormonal rather than a cytokine response. The liver will immediately have a metabolic stress placed on it as it as gluconeogenesis is stimulated in hepatocytes. After a period of time following tissue injury the Kupffer cells of the liver will begin to transcript for cytokines and the liver hepatocytes will be stimulated by these cytokines to manufacture acute phase proteins. Whether the late increase in hepatic arterial flow is a direct effect from cytokines or a reactive response to increased metabolic rate in the liver is not readily addressed by this study.
Finally, we wished to assess changes in blood flow in the intensive care setting. It has previously been reported that total liver blood flow is increased in critical illness compared to controls. Studies over time have suggested that initially there is an increase in total liver blood flow compared to controls; however no significant change after 24 hours. In Chapter 5 hepatic arterial and portal venous blood flows were measured in intensive care patients during their ITU admission. There was no correlation between blood flows and level of systemic inflammatory response as assessed by C-reactive protein. A greater variability in blood flow measurements was seen in non-survivors. The absence of drop in portal flow may be explained by aggressive fluid resuscitation in the first 24 hours of the ITU stay. The absence of an increase in hepatic or total liver blood flow was not clear.
The present thesis identified changes in hepatic arterial and portal venous blood flow between groups of patients and following an inflammatory stimulus. Future work requires to examine which hormones and cytokines modify liver blood flow. This may be assessed by repeating the study in Chapter 4, where liver blood flow was measured following surgery, and by sampling blood for hormonal and cytokine analysis. If the intrinsic control of the liver is responsible for these changes it would be important to show that the liver is more metabolic as the changes occur. An intervention to lower the metabolic demand on the liver is difficult and may not be ethical given that the normal response to injury or infection is an acute phase response. One area where this could be said to have been performed is the tight control of blood glucose in intensive care medicine which reduces the demand on the liver to perform gluconeogenesis. While there are other benefits of lower serum glucose levels this may contribute to the reported improved outcome in such patients.
Clinically, measurement of hepatic arterial and portal venous blood flow has been shown to be feasible in the critically ill patient and may be used as a non-invasive measurement of the liver response to a drug or therapy.
Actions (login required)