Patient and graft survival after pediatric liver transplantation (LT) has improved significantly: children undergoing LT have a patient survival of 79% and graft survival of 63% 10 years post-transplant, according to the European Liver Transplant Registry data . In 2018, a single-center study reported patient and graft survival rates of 83% and 73% at 10 years and 79% and 64% at 20 years after pediatric LT, respectively. However, the long-term outcome of these grafts remains unknown. The Graft survival in both pediatric and adult LT recipients declines in the long term after LT ( Fig. 44.1 ). However, while recurrent liver disease can affect graft survival after adult LT, the etiology of chronic graft loss is less clear in pediatric LT recipients. The importance of chronic graft injury is underscored by the potentially longer life expectancy of pediatric LT recipients. To achieve long-term quality of life for the recipients without the need for retransplantation, maintaining graft integrity and function and minimizing long-term injury is essential.
Immunological, infectious, and late vascular or biliary complications can account for graft injury and late graft dysfunction. Patients can present with or without abnormal liver tests. In this chapter, we focus on chronic subclinical graft injury and discuss differential etiologies of late graft dysfunction. We characterize chronic graft injury, the factors influencing pathophysiology of long-term graft loss, and potential interventions to prevent allograft damage.
Most long-term pediatric LT recipients enjoy a good quality of life without biochemical evidence of allograft dysfunction. However, studies of protocol liver allograft biopsies demonstrated a high prevalence of unexplained inflammation and fibrosis in the biopsies obtained more than 1 year post-LT in asymptomatic children with (near) normal liver biochemistry. These published single-center reports describe heterogeneous cohorts of chronic liver allograft disease in long-term survivors of pediatric LT.
Natural History of Chronic Graft Injury
Histological studies of protocol biopsies obtained from children with normal liver biochemistry up to 10 years after LT have demonstrated graft hepatitis and fibrosis at unexpectedly high rates. Studies of serial biopsies have shown that the prevalence and severity of fibrosis increased over time. The cause for chronic injury was not identified in the majority of patients. Although numerous studies have addressed the problem of chronic, subclinical graft injury both in adults and in children, underlying mechanisms remain unidentified.
Apparently non-immunological causes, such as donor-derived factors, type of graft, technical aspects, and biliary or vascular complications, may potentially contribute to liver allograft fibrosis. Previous studies reported that partial liver grafts, prolonged cold ischemic time, and biliary complications were risk factors for long-term graft fibrosis. Donor age, however, was not an independent risk factor. In contrast, one study showed that technical modifications of the organs, such as split and living donor grafts, were not associated with a higher risk of graft failure, retransplantation, or death. One-year graft survival of split and living donor liver grafts improved significantly in the last 15 years. Consequently, the risk of early and late patient death following split LT is not different from whole graft LT at present modified liver grafts have a significant capacity to regenerate and adapt in response to the metabolic requirements. In a retrospective study, measured volumes of split livers increased by 59% after 5 years and 170% after 10 years.
Technically modified grafts are essential to expand the pool of donor organs, in particular for small children. The propensity of these grafts to early surgical complications, such as bile duct strictures, has repeatedly been discussed. However, the focus has shifted to long-term complications due to accumulating evidence that alloreactivity and humoral response may have an impact on chronic graft injury. Several subsequent studies have shown a high prevalence of unexplained graft inflammation (22%–74%) in the biopsies obtained more than 1 year after LT. There was also a high frequency of liver graft fibrosis (27%–97%) in all of these studies. Two studies showed a strong correlation between the degree of inflammation and a stage of fibrosis, following the concept that graft inflammation contributes to the pathogenesis of post-LT fibrosis. The relevance of humoral response has been supported by positive titers for anti-nuclear antibody, smooth muscle antibody (SMA), elevated immunoglobulin levels, and increased prevalence of angiotensin-2 type 1 receptor antibodies (ATR1), whereas the specific alloimmune response has been reinforced by increased prevalence of donor-specific antibodies (DSAs) directed against human leukocyte antigen (HLA) class II antigens.
Most of the liver recipients with chronic graft injury remain clinically stable in the second decade after pediatric LT. In some patients, however, progressive fibrosis may lead to portal hypertension, graft dysfunction, and retransplantation. Presumably, advanced liver fibrosis can lead to portal hypertension over decades, necessitating retransplantation in a significant proportion of patients, taking into account the longer life span of pediatric liver recipients. A recent retrospective study on histopathology of explanted grafts at retransplantation demonstrated a high prevalence of chronic idiopathic hepatitis and progressive fibrosis in long-term pediatric LT recipients. The authors observed a shift in the histopathology of failed liver grafts over three decades, with an increasing prevalence of chronic idiopathic hepatitis, which was associated with progressive fibrosis and ultimately graft failure.
Histological Patterns of Chronic Graft Injury
Although chronic graft injury may not be apparent in the beginning, distinct early features of chronic liver allograft inflammation and fibrosis have been described.
Hepatitis in Long-Term Liver Graft Recipients
Unexplained chronic graft inflammation (idiopathic posttransplant hepatitis) is characterized by a predominantly portal infiltrate of mononuclear inflammatory cells. Inflammation of bile ducts and portal vessels is typically minimal or absent. Portal inflammation is regularly associated with mild to moderate degrees of interface hepatitis ( Fig. 44.3 ). Lobular inflammatory changes may also be present; they characteristically involve centrilobular regions and may be associated with central perivenulitis (CP).
Scoring systems used to grade chronic hepatitis in the native liver (e.g., METAVIR, Ishak) are only suitable for grading portal and periportal inflammatory changes in the liver allograft biopsies. The severity of CP can be better scored according to the system proposed by the Banff Working Group. Otherwise, chronic graft hepatitis with dominant CP and only minor or even absent portal inflammation may be underestimated. Severe, late CP can be associated with adverse long-term outcomes, even if there is only mild portal inflammation.
Fibrosis in Long-Term Liver Graft Recipients
Three main patterns of liver graft fibrosis were recognized in posttransplant biopsies from pediatric liver allograft recipients: periportal ( Fig. 44.4 ) , perisinusoidal, and centrilobular fibrosis. Preliminary findings suggest different pathogenic mechanisms for these patterns of fibrosis. Periportal fibrosis can develop following initial portal inflammation and interface hepatitis that is present in children with chronic hepatitis ( Figs. 44.4 and 44.5 ). The typical picture of centrilobular fibrosis has been described as a complication of CP in children with chronic rejection (CR). Several studies found evidence to suggest that perisinusoidal fibrosis occurring in the centrilobular region was associated with chronic antibody-mediated rejection (AMR).
Standard scoring systems used to stage fibrosis in the native liver fail in staging the distinctive patterns of fibrosis in pediatric allografts. A specific scoring system focusing on the three patterns of allograft fibrosis, the liver allograft fibrosis (LAF) score, was developed by Venturi and colleagues ( Fig. 44.2 ) . The severity of fibrosis is assessed semiquantitatively in separate histological compartments: around portal tracts, in sinusoids, and around centrilobular veins. These three compartments contribute to a total LAF score. This LAF score helps to describe and quantify stages of allograft fibrosis specifically and is additionally useful to describe dynamics of the fibrosis progression in serial biopsies of pediatric liver transplant recipients.
Differential Diagnoses of Chronic Graft Injury
On detection of liver allograft fibrosis and chronic graft hepatitis several differential diagnoses have to be taken into account.
Recurrent Liver Disease
Recurrent liver disease plays an important role in the management of adult liver transplant recipients. In children, however, recurrence of underlying liver disease is a less prominent complication in the long-term post-LT management. A minority of children and adolescents have received a liver transplant for diseases that may relapse, including immunological (e.g., primary sclerosing cholangitis, autoimmune hepatitis, autoimmune sclerosing cholangitis, and bile salt export pump deficiency) and oncological (e.g., hepatoblastoma and hepatocellular carcinoma) disease. In patients transplanted for these indications, the recurrent liver disease can affect graft function and long-term outcome, so that this potential risk should be considered. The diagnosis of recurrent autoimmune liver disease can be problematic because of the histological overlap with T-cell-mediated rejection and difficulties in interpreting serum autoantibody levels after transplantation. Diagnosis and treatment of recurrent liver disease are described elsewhere (see Chapter 27 ).
Hepatitis E Virus Infection
Hepatitis E virus (HEV) is a recently described cause of chronic graft hepatitis with progressive fibrosis in adults and children. Post-LT, most patients are infected by genotype 3. HEV infection is usually acquired enterally and has been more prevalent in some geographic regions, (e.g., India). More recently, HEV has been identified as an emerging pathogen in developed countries as well, particularly among immunosuppressed solid organ transplant recipients. In those, HEV infection is most likely caused by foodborne transmission. Blood transmission of HEV or even organ-related transmission of HEV has rarely been reported in LT recipients. In the case of transmission, approximately 50% to 80% of adult LT patients infected with HEV develop chronic infection, and 10% to 15% of cases progress to cirrhosis. HEV is also prevalent in pediatric LT recipients, and 8.3% of children are immunoglobulin (Ig) G positive. Cases of chronic HEV infection with progressive allograft hepatopathy have also been observed in pediatric LT recipients. In chronic liver graft dysfunction, HEV RNA test by reverse transcriptase-polymerase chain reaction in blood and/or stool is recommended. The treatment should include modification of immunosuppression and antivirals. Initially, the immunosuppression should be reduced, but if this is not possible, or if there is no HEV RNA clearance within 3 months, treatment with ribavirin (15 mg/kg/day) is recommended for 3 months. Close monitoring of anemia and renal function, as well as HEV load, is required. Ribavirin was shown to be successful in treating HEV infection and preventing graft failure because of acute HEV infection.
Chronic and Antibody-Mediated Rejection
As already mentioned, there is a particular clinical and histopathological overlap between insidious, chronic graft injury, and chronic rejection (CR), or chronic antibody-mediated rejection (AMR). However, many of these patients with apparent or inapparent chronic graft injury do not fulfill the diagnostic criteria for CR or chronic AMR.
CR may itself cause long-term graft dysfunction and fibrosis. Its clinical presentation usually is associated with jaundice, pruritus, elevation of bilirubin, aspartate aminotransferase (AST), alanine aminotransferase, alkaline phosphatase, and gamma-glutamyl transferase levels, which is usually in contrast to the patients with subclinical graft injury. The Banff group defined the histological criteria of CR, including biliary epithelial senescence changes affecting a majority of the bile ducts with or without bile duct loss, foam cell obliterative arteriopathy, and bile duct loss (at least two criteria have to be fulfilled). In their latest update, early and late CR were discriminated, and criteria were more specifically detailed. In a recent study, CR was observed in 37.5% (48 of 128) of 20-year survivors after pediatric LT. In a SPLIT registry study analyzing patients with late graft loss (35 of 872 patients), CR was the cause for 37% of the lost grafts. Some reports have suggested that tacrolimus may markedly reduce CR in comparison with the cyclosporine regimens. Treatment of CR can be challenging, and graft loss is frequent. Of note, rescue treatment with mammalian target of rapamycin (mTOR) inhibitor may be of benefit in some cases.
Chronic AMR of the liver allograft is an emerging entity. The histological picture has a substantial overlap with the silent graft injury caused by idiopathic graft hepatitis. Diagnosis of chronic AMR can be made according to Banff working group criteria as follows:
Histological pattern of injury consistent with chronic AMR, requiring both of the following:
Otherwise unexplained and at least mild mononuclear portal and/or perivenular inflammation with an interface and/or perivenular necroinflammatory activity.
At least moderate portal/periportal, sinusoidal, and/or perivenular fibrosis.
Detection of circulating HLA DSAs in serum samples of patients with subclinical graft injury.
At least focal C4d positivity (> 10% portal tracts, microvascular and endothelial). However, in cases of so-called “possible chronic active AMR” C4d staining can be minimal or absent.
Exclusion of other insults that might cause a similar pattern of graft injury (e.g., hepatitis E infection).
Silent chronic allograft injury lacks specific/typical clinical or biochemical features; many histopathological features observed in protocol biopsies overlap with diagnostic criteria of chronic AMR and/or T-cell-mediated rejection.
Diagnostic and Surveillance of Chronic Graft Injury
Because of the high prevalence of chronic allograft hepatopathy in patients with normal or near-normal biochemical tests, long-term surveillance of pediatric liver transplant recipients has become increasingly important. In addition to patient history and examination, as well as serial laboratory measurements, invasive and noninvasive tests, including imaging and immunological parameters, should be used to monitor liver graft function. Biomarkers of the graft injury are discussed in detail in Chapter 30 .
Protocol Liver Biopsy
Liver biopsies are the gold standard for the evaluation of liver graft integrity. As previous studies have documented, abnormal histopathological findings are common in clinically and biochemically unremarkable patients after pediatric LT. Therefore, the biopsies are important to document the condition of the graft and to identify graft hepatitis or fibrosis more than 1 year after pediatric LT. Furthermore, immunohistochemical biomarkers like the expression of cytoplasmic alpha-smooth muscle actin (ASMA) could be considered for early prediction of the development of graft fibrosis. Assessment of the graft during the process of presumed operational tolerance and weaning chronic immunosuppression are additional reasons to encourage protocol liver biopsies in clinical management.
Protocol liver biopsies may provide relevant information about inflammatory changes or fibrosis. These results, in turn, can lead to modification of immunosuppressive treatment. The benefits of the biopsies as a diagnostic screening tool must be weighed against potential complications. The disadvantages of routine biopsies, including bleeding complications and significant costs, are well known. In current studies after pediatric LT, the rate of major complications (hemorrhage requiring blood transfusions) was lower than 1% with no mortality. In a large retrospective study of 68,276 percutaneous liver biopsies in adults, the risk of mortality was also low (0.9/10,000 biopsies). It is essential to identify patients with an increased risk of complications. Currently, protocol biopsies are the only available tool to assess subclinical chronic graft injury. However, uncertainties related to the clinical significance and a lack of guidance for therapeutic interventions require individualized consideration for the protocol biopsy. Further evidence of the long-term outcome of chronic allograft injury and the effectiveness of consequent treatment strategies is required before definite general recommendations can be made.
In the context of minimization and withdrawal of immunosuppression, however, the ESPGHAN Hepatology Committee and the Banff Working Group recommend a liver biopsy before and after modifying immunosuppression. Liver histology is necessary to discriminate patients under increased risk for allograft rejection. Monitoring of blood tests, liver ultrasonography, or noninvasive fibrosis markers are not considered sufficiently sensitive in assessing chronic damage to the liver allograft.
Noninvasive Assessment of Liver Graft Integrity
In the light of the risk for post-biopsy complications, including hemorrhage and the requirements for sedation/anesthesia, several noninvasive tests for assessment of chronic graft injury have been proposed. Apart from HLA and non-HLA antibodies, serological markers of liver transplant fibrosis were suggested (see Chapter 30 ). Furthermore, imaging studies including different modes of ultrasound and magnetic resonance elastography have been explored for assessment of liver fibrosis in transplant and non-transplant patients.
Dynamic ultrasound elastography is based on the assessment of the propagation of shear waves within tissues to calculate the viscoelastic properties. Transient elastography (FibroScan) is a first-generation dynamic ultrasound elastography method and does not offer a conventional B-mode image. It relies on the application of a short external shear wave pulse that is tracked by using one-dimensional ultrasound imaging. The test can be applied rapidly and repeatedly, and children tolerate it well. However, obesity, ascites, or limited operator experience can impair results in up to 20% of cases. Its diagnostic accuracy is better for higher-degree fibrosis/cirrhosis than for mild or moderate fibrosis. The technique has been evaluated in adult LT recipients showing excellent results for the detection of advanced fibrosis/cirrhosis (area under the receiver operating characteristic curve [AUROC], 0.86–0.98), but transient elastography was less useful for mild fibrosis (AUROC, 0.81). In pediatric LT, small studies aiming at correlating transient elastography to post-LT fibrosis have confirmed the potential of the technique to detect advanced fibrosis. Assessment of split or reduced grafts can be difficult because of a lack of defined cutoff values.
Acoustic radiation force imaging (ARFI) belongs to the second generation of ultrasound elastography methods measuring shear wave speed while offering conventional ultrasound B-mode imaging. Thereby, stiffness can be measured in a selected region of interest, making the assessment feasible in almost all patient groups. Comparable to transient elastography, ARFI accurately detected advanced fibrosis (AUROC, 0.98) in children and adults with chronic liver disease, but ARFI was less compelling in detecting mild fibrosis (AUROC, 0.83). Small studies in pediatric liver transplant recipients showed a good correlation of liver stiffness measurements by ARFI imaging with liver graft fibrosis. ARFI had good accuracy for diagnosing significant portal or moderate perisinusoidal graft fibrosis after pediatric living donor LT.
Magnetic resonance elastography (MRE) is a highly accurate, noninvasive technique for diagnosis and staging of liver fibrosis in the LT recipients, independent of body mass index and degree of inflammation. MRE has already been validated for the detection of liver fibrosis, including mild to moderate fibrosis in adult LT recipients and chronic pediatric liver disease. MRE may also be promising in pediatric LT recipients, but requires a prospective evaluation. Disadvantages of MRE include costs and the need for anesthesia in small children.
AST-to-platelet ratio index (APRI) is a straightforward, noninvasive score to assess liver fibrosis, which has been well validated for hepatic fibrosis in hepatitis C. In pediatric LT recipients, a higher APRI was associated with significant fibrosis (Ishak score ≥ 3; AUROC, 0.74). However, other studies did not show a correlation between APRI and stage of fibrosis. Of note, most of the patients with chronic graft injury initially show normal biochemistry despite advanced graft fibrosis.
In conclusion, noninvasive investigations, including imaging studies and biomarkers, can help monitor progressive fibrosis over time. Nevertheless, these tests cannot substitute the protocol liver biopsies entirely at this stage because they cannot identify patients who exhibit early stages of graft fibrosis and inflammation, pointing to their increased risk for progressive graft injury and graft loss.
Donor-Specific HLA Antibodies
In the last decade, donor-specific HLA antibodies (DSAs) and chronic AMR have been identified as important risk factors for long-term allograft dysfunction and chronic allograft failure in pediatric LT (see Chapter 17 ). A meta-analysis showed that the prevalence of DSAs was 41% after pediatric LT. Risk factors for the emergence of de novo DSAs include low trough levels of calcineurin inhibitors, cyclosporine versus tacrolimus use, younger age, and previous transplants.
Although the pathogenic role of DSAs in LT is still debatable, evidence is accumulating that preformed DSAs with high mean fluorescence intensity (MFI) values and de novo DSAs developing post-transplant are associated with chronic graft injury. Higher DSAs levels with MFI values over 5000 increase the risk of associated graft injury. There is emerging data suggesting that subsets of DSAs—those of the IgG3 subclass and/or those that bind the first subcomponent of complement complex C1 (C1q)—are particularly damaging. A recent study demonstrated an association between HLA class II DSAs and portal inflammation, which, over time, drives progressive fibrosis in the portal areas. However, it remains unclear whether DSAs are directly responsible for the chronic graft inflammation or are instead just a surrogate marker of chronic alloimmune recognition and injury of the graft. Importantly, severe graft injury may also occur with negative DSAs.
Complement activation in the liver tissue plays a significant role in the pathogenesis of AMR. The deposition of C4d in the portal microvascular endothelium with extension into inlet venules or periportal sinusoids in severe cases is the histochemical correlate of humoral graft inflammation in the liver tissue. This suggestion of antibody-mediated subclinical inflammation is more conceivable, as the C4d deposition in the hepatic tissue has been shown to correlate with the antibody presence. Therefore, C4d staining is an additional essential marker of antibody mediation in chronic liver graft injury. C4d positivity could be associated with chronic graft damage with or without a clinical picture of AMR. However, nonspecific staining is possible, and negative staining does not exclude antibody-mediated injury.
Different methods of analysis have been used for testing HLA antibodies, making interpretation and their comparison difficult. An isolated MFI value can be misleading. Therefore, prospective longitudinal studies of DSAs following pediatric transplantation are needed to assess the emergence and dynamics of DSA positivity to correlate the DSAs presence with clinical conditions and to draw conclusions relevant to clinical management.
In 40% to 60% of the liver allograft recipients, liver autoantibodies (usually anti-nuclear and smooth muscle actin antibodies) were detected by 5 to 10 years after LT. In most studies, the autoantibody positivity was associated with chronic graft inflammation. Liver autoantibodies, however, appear not to be the cause of the graft injury, but rather a reflection of its damage and could potentially be a helpful marker for chronic allograft hepatitis. Liver autoantibodies, de novo DSAs, and antibodies to glutathione S-transferase T1 are associated with plasma cell–rich rejection, also known as de novo autoimmune hepatitis. This entity is an uncommon cause of late graft dysfunction. The histopathological picture is very typical, showing portal and perivenular plasma cell infiltrates with interface activity. In pediatric LT recipients, de novo autoimmune hepatitis or plasma cell-rich rejection was found in about 2% to 7% of post-LT patients. This type of graft injury needs modified immunosuppression and can be associated with progressive graft dysfunction and the risk of graft failure (see Chapter 27 ).
More recently, non-hepatocyte autoantibodies were suggested to play a role in chronic graft injury. Detection of anti-ATR1 and anti-endothelin-1 type A receptor (ETAR) autoantibodies was associated with an increased risk of rejection and progression of liver allograft fibrosis in adult patients. Similarly, in pediatric transplant recipients, anti-ATR1 autoantibodies were associated with advanced fibrosis. However, the role and pathogenic mechanism of anti-ATR1 and anti-ETAR autoantibodies in chronic graft injury need further evaluation.
Chronic Graft Injury: Etiology Still Unclear?
The etiology of chronic graft injury is not fully understood. The surgical complications can play a role in some cases of nonprogressive graft fibrosis (see above: Natural History of Chronic Graft Injury). However, there is increasing evidence that a distinct form of subclinical alloimmune reaction could significantly contribute to evolving inflammation and fibrosis. This assumption could particularly apply to the pediatric population, in whom the recurrence of underlying liver disease as a cause of late graft dysfunction is exceptional. Furthermore, so-called “idiopathic post-transplant hepatitis” may represent a form of subclinical late rejection. Evidence supporting this hypothesis includes the following:
The frequent presence of autoantibodies in patients with chronic graft hepatitis and fibrosis.
There is a strong association between HLA class II DSAs and chronic graft hepatitis and fibrosis. The chronic graft injury correlates with a higher prevalence and a higher MFI value of DSAs. DSAs are mainly associated with perisinusoidal fibrosis, suggesting a unique mechanism for this pattern of fibrosis.
Previous episodes of T-cell-mediated rejection are a risk factor for chronic graft hepatitis and fibrosis. This observation again implicates alloimmune mechanisms.
Histological features of chronic graft hepatitis overlap with diagnostic criteria of CR or chronic AMR
Recent studies in adult and pediatric LT showed that portal inflammation with interface activity and varying degrees of fibrosis were associated with specific transcriptional patterns: interferon-γ-regulated gene signatures associated with T-cell-mediated rejection were significantly overexpressed. Overall, the transcriptomic profile analyzed in the graft tissue affected by portal inflammation and interface activity resembles that of T-cell-mediated rejection.
Based on the hypothesis that chronic graft injury is an alloimmune response and a type of subclinical chronic transplant rejection, a treatment modification through intensification of immunosuppression appears to be justified upon the detection of such injury. However, the guidelines for these treatment modifications remain not well defined. Some groups suggest that the graft inflammation and, less convincingly, graft fibrosis could improve with an intensification of immunosuppression. In a single-center study, for example, steroids were added to the therapy regimen after detecting significant hepatitis and fibrosis in the protocol biopsies after LT in association with positive serum liver autoantibodies. In follow-up biopsies after 5 years, the hepatitis activity had decreased from 43% to 26%. The effect on fibrosis remained uncertain. However, randomized controlled interventional studies on the treatment of chronic graft injury are lacking. The understanding of long-term effects and outcome of enhanced immunosuppressive treatment in graft hepatitis need to be improved before universal therapeutic recommendations can be made.
Conclusion and Further Concepts
Liver allografts are histologically abnormal in the majority of patients surviving long-term following LT in childhood. These abnormalities, including subclinical graft hepatitis and fibrosis of different degrees, have been observed in protocol biopsies from patients who appear to be symptom-free with biochemically normal graft function. Nevertheless, there is emerging evidence to suggest that this chronic injury could have an important impact on long-term graft survival. Studies of serial biopsies have shown that the occurrence and severity of fibrosis could increase with the time and, in some cases, could lead to graft loss. It is possible that some of these changes are the consequences of prolonged insidious inflammatory reactions related to a subclinical alloimmune injury. Recent data suggested that chronic graft injury is most likely a mixed immunological process influenced by both humoral sensitization and T-cell alloreactivity. These findings support the two-hit hypothesis where the presence of DSAs does not necessarily independently lead to chronic graft injury. Contributors that enhance alloreactivity are ischemia-reperfusion injury, recurrent episodes of T-cell-mediated rejection, and fluctuating levels of immune suppression because of patient nonadherence (see Fig. 44.6 ).