Kidney transplant outcomes in HIV-positive patients: a systematic review and meta-analysis

Background Kidney transplantation is now a viable alternative to dialysis in HIV-positive patients who achieve good immunovirological control with the currently available antiretroviral therapy regimens. This systematic review and meta-analysis investigate the published evidence of outcome and risk of kidney transplantation in HIV-positive patients following the PRISMA guidelines. Methods Searches of PubMed, the Cochrane Library and EMBASE identified 27 cohort studies and 1670 case series evaluating the survival of HIV-positive kidney transplant patients published between July 2003 and May 2018. The regimens for induction, maintenance therapy and highly active antiretroviral therapy, acute rejection, patient and graft survival, CD4 count and infectious complications were recorded. We evaluated the patient survival and graft survival at 1 and 3 years respectively, acute rejection rate and also other infectious complications by using a random-effects analysis. Results At 1 year, patient survival was 0.97 (95% CI 0.95; 0.98), graft survival was 0.91 (95% CI 0.88; 0.94), acute rejection was 0.33 (95% CI 0.28; 0.38), and infectious complications was 0.41 (95% CI 0.34; 0.50), and at 3 years, patient survival was 0.94 (95% CI 0.90; 0.97) and graft survival was 0.81 (95% CI 0.74; 0.87). Conclusions With careful selection and evaluation, kidney transplantation can be performed with good outcomes in HIV-positive patients.


Background
Traditionally, human immunodeficiency virus (HIV)positive patients (HIV+) has not been considered to be good candidates for solid-organ transplantation for the poor prognosis of HIV patients. However, with the introduction of antiretroviral combination therapy (cART), the survival of HIV+ patients have been great improved. While the frequency of Acquired Immune Deficiency Syndrome (AIDS)-related events has consequently decreased, mortality due to organ failure has become a significant concern.
The initial attempts at kidney transplantation (KT) in HIV+ patients led to poor outcomes, but better results occurred with the availability of highly active antiretroviral therapy (HAART) [1,2].
In this scenario, KT started to be proposed as a treatment even as "standard-of-care" for end-stage renal disease (ESRD) in selected HIV+ patients [3].
A multicentre study in the USA found that the survival rates for HIV+ recipients fall between those reported for older KT recipients and for all recipients in the American national database [4].
Despite these encouraging results, many issues still need to be addressed. Among the more relevant are the elevated incidence of acute rejection (AR), lower patient survival (PS) and graft survival (GS), and the hurdles caused by the interaction of immunosuppressive and antiretroviral (ARV) drugs. We conducted a systemic review and meta-analysis to determine the effectiveness of KT in the presence of HIV. Specifically, we examined PS and GS, AR and infectious complications in HIV+ patients who have undergone KT. Zheng et al. AIDS Res Ther (2019) 16:37

Study design
The study design of a systematic review and metaanalysis was chosen to define the published evidence of the effectiveness of KT in HIV+ patients. The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement standards [5]. Our review was registered at the International Prospective Register of Systematic Reviews (PROSPERO CRD42018109178).

Search strategy
We searched the Medline (1966 to June 2018), EMBASE (1974 to June 2018), and Cochrane Controlled Trials Register databases to identify studies that referred to KT in HIV+ patients; we also searched the reference lists of the retrieved studies. The following search terms were used: KT, HIV+, AIDS. A combination of subject headings and keywords for KT, HAART, HIV+ recipient, allograft survival, antiretroviral therapy, donor selection, ESRD and immunosuppression was used for the literature search.

Eligibility criteria
Cohort studies and case-control studies were all eligible for inclusion if they reported outcomes of KT in HIV+ patients. Studies reporting outcomes shorter than 12 months post transplantation and transplantation occurring before HAART were introduced were excluded. Articles were independently assessed by 2 reviewers (X Z and WR X) according to the predetermined eligibility criteria. Any disagreement between reviewers was resolved by discussion with a third reviewer (XP H).

Data extraction
All data were extracted independently by 2 reviewers (X Z and WR X) onto a Microsoft Excel spreadsheet (XP Professional Edition; Microsoft Corp, Redmond, WA), and any discrepancies were resolved by consensus. The following information was collected for each study: the study country, sample size, inclusion criteria, exclusion criteria, induction and maintenance immunosuppression, HAART regimen, mean CD4 T-cell counts (CD4 counts) pre-transplant and post-transplant, infectious complications, post-transplant neoplasia, PS and GS at 1 and 3 years, and AR rate. In order to analyse data of Infectious complications (IC), all infections requiring hospitalization were registered.

Quality grading of studies
The quality of each study used for the meta-analysis was assessed based on the Newcastle-Ottawa-Scale (NOS) for cohort studies [6]. The evaluation of study quality included the following three categories: (i) selection (4 items), (ii) comparability (2 items), and (iii) the assessment of outcome (3 items). The NOS ranges from zero to a maximum of 9 points. Five authors (X Z, W X, S Z, Y X and Y Z) independently assessed the articles. The overall NOS score was determined as the median of all 5 individual NOS assessments. Study quality was graded as good (≥ 8 points), fair (6 or 7 points), and poor (≤ 5 points) [6].

Data analysis
We undertook the descriptive analyses to identify the number of studies with relevant data, the countries where the studies were conducted, and other population attributes. The transplant outcome data were pooled using different transformations according to their different normal distribution conditions.
The data for PS, GS, AR, IC at 1 year and PS at 3 years were analysed using log transformation.
The data for GS at 3 years were analysed using arcsine transformation.
The transformed data were combined to estimate the pooled percentages with 95% confidence intervals using a random-effects model.
The transformed data were combined to estimate the pooled percentages with 95% confidence intervals using a random effects model [7] and presented as forest plots. We assessed the heterogeneity among studies using the Cochran Q test ( χ 2 n−1 ; p < 0.05 to denote statistical significance) and estimated the amount of variation by I 2 [8]. Statistical sources of heterogeneity were explored by examining the relationship between one or more study-level characteristics and the effect sizes that were observed in the studies using weighted least squares meta-regression. A rank correlation test of funnel plot asymmetry (z) was used to assess the presence of publication bias. Statistical analysis was performed using the R statistical software package (R Development Core Team, Vienna, Austria; URL: http://www.R-proje ct.org; version 2.9.0), using the software libraries 'meta' and 'metafor' , for the meta-analysis and meta-regression models, respectively.

Systematic study review
The search strategy identified 86 citations (Fig. 1), and among these, we identified 53 studies that appeared to bge relevant to our study. Finally, 27 of these studies, containing 1670 cases, met the inclusion criteria. Agreement between reviewers for assessment of study eligibility was 100%.
Detailed characteristics of all included studies are provided in Table 1. A majority of the studies were conducted in the US or Europe. All these details are summarized in Tables 2 and 3. Regarding immunosuppression and rejection, most of the patients received antibody induction therapy with different regimens containing basiliximab, daclizumab, antithymocyte globulin (ATG) or methylprednisolone. The maintenance regimens were mainly composed of cyclosporin A (CSA), mycophenolate mofetil (MMF), tacrolimus (TAC) and steroids, which were the same as the maintenance regimens for HIV-negative patients.

Quality of studies included in the meta-analysis
Each of the 27 studies included in the meta-analysis was assessed by the NOS to investigate the risk of bias within the studies. Table 4 shows the results of the quality assessment. None of the studies had less than three points in the category selection. Two studies controlled for age and gender, and 9 controlled for other factors, such as the HAART regimen and/or immunosuppression therapy. Finally, 5 studies were graded as good quality and 22 as fair quality.

Acute rejection post KT
Twenty-five studies including 1051 patients reported AR post KT at 1 year, and the results of the analysis are shown in Fig. 6. At 1 year, 33% (95% CI 0.28; 0.38, I 2 = 60%) of patients had AR.

Infectious complications post KT
Nineteen studies including 584 patients reported IC post KT at 1 year; the results of the analysis are shown in Fig. 7. At 1 year, 41% (95% CI 0.34; 0.50, I 2 = 59%) of patients had IC.

Discussion
To our knowledge, this is the first systematic review and meta-analysis of such a large scale to report the outcomes of KT in HIV+ patients. We review and meta-analysis the outcomes in HIV+ KT patients, and looks at the 1and 3-year GS/PS and AR rate.
The availability of cART has made KT a feasible treatment for selected HIV+ patients with ESRD, with outcomes somewhat inferior to those observed among the overall population of KT recipients [4,[9][10][11].

Outcomes of KT
KT is now a viable treatment for select patients with HIV and ESRD. Moreover, the high incidence of morbidity and mortality resulting from cardiovascular issues in HIV+ patients [12,13], as well as the negative effects of prolonged steroid use on conditions associated with cardiovascular risk, such as diabetes, dyslipidaemia, and hypertension, are well known [14,15].
However, data regarding long-term outcomes and comparisons with appropriately matched HIV− patients are still lacking.
Locke et al. analysed 510 adult KT recipients with HIV matched 1:10 with HIV− controls. They found that HIV− and HIV mono-infected KT recipients had similar GS and PS, whereas HIV/HCV co-infected recipients had worse outcomes [16].
Izzo et al. found that the survival rate of patients was 82.1% and functioning grafts was 71.4% [17], and a recent report from the Italian national transplantation registry showed a PS rate of 95% and a GS rate of 85% between 2006 and 2014 [18].
Stock et al. [4] reported a survival rate of 94.6% 1 year after transplantation (88.2% after 3 years) in a multicentric trial (150 patients), and in a published review with a small number of patients, the survival rate was 93% within the first year of transplantation (254 patients) [19]. What's more, as the high incidence of co-infection with HCV in HIV+ patients, co-infection is likely a driver of poor outcomes [20].

Immunosuppression therapy
One of the most challenging goals in solid-organ transplantation is to tailor the immunosuppressive regimen for each individual patient to minimize immunosuppression while still preventing AR. Opportunistic infections  and malignancies often attributed to immunosuppression itself remain a significant cause of death after transplantation. In the field of HIV+ organ transplantation, finding a balanced approach to immunosuppression is even more critical. Currently, the vast majority of KT patients receive induction immunosuppression, which has been shown to greatly reduce the risk of rejection and improve PS and GS [21]. As shown in our analyses, most of the HIV+ KT patients received induction therapy. The two most commonly used induction agents are ATG and IL-2 receptor blocker (anti-IL2R) [22].
Guidelines from the Kidney Disease: Improving Global Outcomes transplant working group recommended anti-IL2R as the first-line treatment for patients at low risk for rejection and ATG for those at high risk [23].
Despite being the standard of care for most HIV patients, the use of induction immunosuppression for HIV+ patients, particularly ATG, remains controversial.
On the one hand, HIV+ patients have high rates of rejection and thus stand to benefit significantly from induction. On the other hand, the risks posed from prolonged lymphocyte depletion are of major concern given that HIV+ patients are perceived to already threaten T cell populations and reduced immunity, both states that are associated with an increased risk of opportunistic infections.
A recent study showed that ATG induction was associated with long-term impairment of T cell function and related infections, even after the patients had normalized CD4 counts [24]. This finding is also a reminder that the CD4 counts incompletely assesses the recovery of an immunocompetent CD4 T cell pool.
The incidence and severity of IC following transplantation are largely dictated by the recipient's capacity for immune reconstitution. A study by Suarez et al. indicate that ATG-induced CD4 lymphopenia can be prolonged, and even at 1 year post transplant, a substantial proportion of patients has CD4 counts < 200/μL [25]. The baseline CD4 counts did not influence the risk of death, graft loss or AR. These findings suggest that although in current practice, HIV+ candidates with pre-transplant CD4 counts between 200 and 349/μL are eligible for KT [26] and are likely to have outcomes similar to those with higher counts, this group of patients carries a substantial risk of lymphopenia and associated infections following ATG induction.  A study by Bossini et al. showed that in HIV+ KT recipients treated with basiliximab and maintained on a calcineurin inhibitor (CNI)-mycophenolic acid (MPA)based regimen, early corticosteroid withdrawal was associated with a very high incidence of AR and that kidney function was worse in patients with rejection [27].
However, in contrast to the studies above, in a large national cohort of 830 HIV+ KT recipients, Kucirka found wide variation in the use of induction immunosuppression, with > 30% of HIV+ patients receiving no induction compared with only 20% of their HIV counterparts. Therefore, the study indicated that the use of induction, including the lymphocyte-depleting agent ATG, was not associated with an increased risk of infections. Despite the fact that induction recipients were at higher risk for AR, the researchers observed lower rates of delayed graft function (DGF), AR, graft loss and days spent hospitalized in the first year after KT as well as a trend towards lower mortality. They suggested that the benefits of induction immunosuppression to prevent graft rejection in HIV+ KT recipients far outweigh the perceived risk of increased infections. Because the study had the largest sample size to date and the cohort was nationally representative rather than a select study population, this study claims to be more credible. Furthermore, the authors accounted for confounding and treatment selection bias, which previous studies did not do, and they did this using inverse probability of treatment weighting (IPTW), a method that allowed them to adjust for many clinical and demographic factors even when modelling relatively rare outcomes such as graft loss and death [28].

Acute rejection
A high risk of AR is a well-known concern in HIVinfected kidney graft recipients. With regard to rejection, most studies observed a higher number of events in HIV+ patients than in HIV− patients. AR may occur as a result of immune dysregulation and the continuous inflammatory state of HIV+ recipients, in whom immunogenicity is increased following allograft implantation [4,29].
Vicari et al. [30] evaluated the outcomes of KT in recipients with HIV infection under HAART in Brazil. The main results showed that HIV+ recipients presented a higher incidence of DGF, rejection, and bacterial infections and had lower PS and GS rates in comparison with a paired control group.
In the study, the incidence of treated AR was higher in the HIV+ group, and the incidence of biopsy-confirmed AR was numerically higher in this group. Additionally, even though an identical incidence of antibody-mediated AR occurred, the incidence of steroid-resistant rejection was numerically higher in the HIV group. Many reports Fig. 7 Pooled estimated proportion of infectious complication at the first year, analyzed using a random effects model have revealed an elevated incidence of AR in HIV+ recipients, varying between 31% and 55% [4,9,10], although a significantly lower incidence was reported in one study [31]. Stock et al. [4] reported that a significant proportion of acute cellular rejections were steroid resistant and that no episodes of antibody-mediated AR were observed in their cohort.
However, Malat et al. [32] described an elevated incidence of mixed cellular and antibody-mediated rejections. Furthermore, Locke et al. [33] reported that HIV+ patients who received ATG induction therapy had a much lower risk of rejection compared to patients without induction and that the risk was similar to uninfected controls.
Gathogo et al. [34] reported that TAC has an impact in reducing the incidence of AR in HIV+ recipients compared to cyclosporin A (CSA). The reasons for such an elevated incidence and severity of AR in HIV+ KT recipients are not clear. Dysregulation of the immune system along with a continuous inflammatory state caused by HIV infection, perhaps in association with a variability in drug exposure, has been hypothesized to explain these almost uniformly elevated incidences of rejection [4,10,35].
In addition, the elevated incidence of acute cellular rejection has been recently hypothesized to partially occur a result of an infiltration of inflammatory cells that occurs in response to tubular cell infection by HIV [36].
A study by Malat showed a relatively higher incidence of mixed rejection in HIV+ recipients compared with that reported for non-HIV transplant recipients. A donor terminal serum creatinine greater than 2.5 mg/dL predicted mixed rejection and was associated with poor outcomes. Donor selection and optimization of immunosuppression may be critical in these patients [36]. Even if rejection was controlled successfully with steroid therapy, these results, as previously reported, suggest a possible scenario where the immune system, damaged by HIV infection, has a worse response to immunosuppressive treatment with respect to the general population, even in patients without a severe immunological dysfunction at the time of transplantation. In our analyses, AR at 1 year was 0.33 (95% CI 0.28; 0.38).

Infectious complications
During the first decades of the renal transplantation era, a serious IC developed in up to 70% of patients following transplantation, resulting in fatal outcomes in as many as 11% to 40% of cases [37]. In a recent case-control study with a median follow-up of 5 years, Ailioaie et al. [38] found a similar incidence of post-transplant IC in HIV+ KT recipients compared with matched KT HIV− controls. An IC incidence of 29% after transplantation was previously reported [19], and the incidence of posttransplant neoplasms has been described as similar to the incidence in HIV− patients. In our analyses, the incidence of IC observed at 1 year was 42% (95% CI 0.34; 0.50, I 2 = 59%), and the rate of incidence of IC observed in this study in HIV+ KT patients is in line with the frequencies reported in a study by Stock et al. [4] where 38% of 150 HIV− KT recipients had at least one infection that required hospitalization.
However, the long-term patient and graft outcome of the whole cohort were not influenced by HIV status but were adversely influenced by infections, as survival was diminished in patients having at least one infection.
Furthermore, one-third of HIV+ KT recipients in a study by Ailioaie et al. did not have any episodes of infection, and repeated infections were not frequent. More importantly, the rate of incident infections was not different between the HIV+ and HIV− matched groups.

Drug interaction
As experience with transplantation in HIV+ patients grow, significant drug-drug interactions between ART and maintenance immunosuppression have been identified as a major clinical challenge.
Post-transplant management of HIV infection with protease inhibitor (PI) and nonnucleoside reverse transcriptase inhibitor (NNRTI)-based ART is complicated by reciprocal drug interactions with immunosuppressive therapy, especially CNI, because of inhibition or induction of P450 cytochrome enzymes. Co-administration of PIs with CNIs requires significantly decreased CNI doses and prolonged dosing intervals to avoid supratherapeutic trough levels.
Despite appropriate CNI dose adjustments, variations of drug serum levels are difficult to control and have been linked to increased graft rejection in HIV+ KT recipients [34,39].
In a study of 150 HIV+ KT patients, the largest to date, higher-than-expected rates of rejection were reported (31% and 41% at 1 and 3 years, respectively) [4]. The authors speculated that increased rates of rejection may have been secondary to altered CNI levels since only one-third of patients on PI-or NNRTI-based regimens underwent CNI dose adjustments.
The increased risk of AR in HIV+ individuals has been largely attributed to reduced exposure to immunosuppressive agents due to drug-drug interactions with ART [4,41,42]. Other factors, such as infection of the allograft, previous alloimmunization and immune activation, might also play roles in predisposition to rejection [43].
This observation might be due to the effects of PI on tacrolimus levels, considering that the overwhelming majority of these patients were on a PI-containing regimen and that more than half had tacrolimus levels above target at the time of infection. PI could also influence the net state of immunosuppression by increasing the level or effect of other immunosuppressants, such as prednisone and mycophenolate.
The most important finding in the present study is the association between PI use and adverse outcomes, namely, reduced 3-year PS and GS, and increased risk of serious non-opportunistic infections. These observations remained true in analyses restricted to patients receiving nucleoside reverse transcriptase inhibitor (NRTI) "backbone"; thus, even after excluding the potential influence of other agents included in the ART regimen, PI continued to be associated with poor outcomes.
However, the use of NNRTI or tenofovir disoproxil fumarate (TDF) did not influence GS. Tenofovir alafenamide (TAF) is a new formulation of tenofovir associated with less kidney (and bone) toxicity [44]. Whether TAF has added clinical benefit over TDF in KT recipients remains to be established.
In a large single-centre study of HIV+ KT recipients conducted by Boyle et al. [45], treatment with TDF at the time of transplant was not associated with 36-month death-censored primary allograft loss after adjustment for DGF and a propensity score for TDF exposure.
Given that specific recipient characteristics, such as hepatitis B co-infection and certain HIV mutations, continue to make TDF-based regimens the most likely to provide adequate viral suppression post-transplant, despite observational data for nephrotoxicity in the nontransplant population.
However, given the limitations of this study, TDF should be reserved for patients who have limited ART options and should be used very cautiously in the KT population, with appropriate dose adjustment and surveillance of kidney function, including kidney biopsy when indicated. Substituting TAF for TDF in KT patients is reasonable, but it should be noted that no data are yet available on long-term kidney outcomes with TAF in KT and non-KT recipients in the setting of both preserved and reduced glomerular filtration rate (GFR).
Since their introduction in 2007, integrase strand transfer inhibitors (INSTIs) have been proposed as preferred post-transplant ART because of a favourable pharmacologic profile with decreased potential for drug interactions [3,4,42,46]. In a study by Stock et al. [4] the majority of patients were on PIs or NNRTIs with only 4% of participants receiving INSTIs; these patients were also receiving PI, NNRTI or maraviroc, making it impossible to draw conclusions about INSTI-based therapy. In a series of 27 HIV+ KT patients in France predominantly on PI or NNRTI-based regimens (93%), 70% required post-transplant ART modification due to drug interactions with CNIs [32].
Recently, Alfano et al., reported that preferred drug included raltegravir and dolutegravir for INSTI class, maraviroc for CCR5 receptor antagonist, lamivudine for NRTI, and rilpivirine for NNRTI, which offered advantage of having no drug interactions [47].
In summary, we believe that INSTI or CCR5-based therapy should be the preferred ART in patients with HIV who undergo KT, primarily because of decreased drug-drug interactions with immunosuppressive medications such as CNIs, enabling easier monitoring of immunosuppressive medications and superior graft outcomes. However, larger and more controlled trials are needed to better assess the long-term outcomes of INSTI-based therapy to elucidate factors related to GS other than direct reciprocal drug interactions.

Conclusions
In conclusion, this systematic review and meta-analysis demonstrated that with careful selection of patients and multidisciplinary evaluation, KT can be performed with good outcome in HIV+ patients. Moreover, with the advent of INSTI-based cART regimens, drug-drug interactions between cART and immunosuppressants have been dramatically reduced. Nevertheless, further studies are needed to optimize immunosuppressive therapy regimens for HIV+ patients, with the aim of reducing the high rate of AR after transplantation. Furthermore, this review still has its limitations, such as lack of sufficient studies, possibility of some overlapping patient cohorts, short of comparator. And we are also looking forward to other novel papers as more and more studies regarding KT of HIV+ patients.