Outcomes in Transplant Recipients Treated with Foscarnet for Ganciclovir-Resistant or Refractory Cytomegalovirus Infection

Abstract

Background

Antiviral-resistant or refractory CMV infection is challenging, and salvage therapies, foscarnet and cidofovir, have significant toxicities. Several investigational anti-CMV agents are under development, but more information is needed on outcomes of current treatments, to facilitate clinical trial design for new drugs.

Methods

Records of solid organ (SOT) and hematopoietic cell transplant (HCT) recipients at a single center over a 10-year period were reviewed retrospectively to characterize those who had received foscarnet treatment for ganciclovir-resistant or refractory CMV infection. Data were collected on virologic responses, mortality, and nephrotoxicity.

Results

Of 39 patients (22 SOT, 17 HCT), 15 had documented ganciclovir resistance mutations and 11/39 (28%) had tissue-invasive CMV. Median duration of foscarnet was 32 days. Virologic failure occurred in 13/39 (33%) and relapses of viremia occurred in 31%. Mortality was 12/39 (31%) and was higher in HCT than SOT (p=0.001), although ganciclovir resistance was more common in SOT (p = 0.003). Doses of ganciclovir or valganciclovir were low in 10/39 (26%) at some time prior to switching to foscarnet. Renal dysfunction occurred in 20/39 (51%) by end of treatment and in 7/25 (28%) after 6 months.

Conclusions

Outcomes of existing treatment for ganciclovir-resistant or refractory CMV are suboptimal, in terms of virologic clearance, renal dysfunction, and mortality. These data should provide background information for future clinical trials of newer antiviral agents.

INTRODUCTION

Ganciclovir-resistant or refractory CMV infection in transplant recipients is frequently characterized by a protracted disease course, organ dysfunction, drug toxicities, recurrences, and high mortality risk^1–3. In addition to patients with genotypically demonstrable CMV resistance mutations, there is widespread recognition that a “refractory” group exists, comprised of patients who do not respond clinically and/or virologically to standard treatment, but who do not have resistance mutations identified on CMV genotyping. The optimal treatment of these infections is controversial. Second-line antiviral medications (foscarnet and cidofovir) carry high risk for adverse effects, including nephrotoxicity (both foscarnet and cidofovir), electrolyte abnormalities (foscarnet), genitourinary ulcerations (foscarnet), and uveitis (cidofovir). High-dose ganciclovir⁴, or combinations of ganciclovir and foscarnet^5–6, are variably effective. CMV immune globulin (CMVIg) is frequently added, although there is little evidence for efficacy in treatment of active infection^7–8. Leflunomide has been used off-label for CMV treatment, but is not always effective⁹. Thus, there is a distinct need for effective antiviral therapies with improved side effect profiles for treatment of resistant or refractory CMV.

There are several investigational agents in development: maribavir¹⁰, brincidofovir (CMX001)¹¹, letermovir^12–13, and artesunate^14–15. However, none of these is currently available for compassionate use.

In designing clinical trials to test the efficacy and safety of investigational anti-CMV agents, additional data are needed to establish baseline response rates and outcomes of treatment with currently available therapies, especially foscarnet. Despite an extensive literature on CMV in transplantation, only a few studies in the last 15 years have described foscarnet treatment outcomes outside of preemptive therapy^16–23. Here, outcomes were reviewed from SOT and HCT recipients who had received foscarnet for treatment of resistant or refractory CMV over a 10-year period at a single center.

MATERIALS AND METHODS

Patients

After approval was received from the Johns Hopkins Institutional Review Board, a list of all inpatients who had received foscarnet at Johns Hopkins Hospital between 1/2005 and 5/2015 was obtained from Johns Hopkins Pharmacy records. This list was cross-checked with a list of patients who had undergone UL97 and UL54 resistance testing in the Johns Hopkins Medical Microbiology Laboratory. The following categories were excluded: patients who had not received a transplant, patients who received foscarnet for viral infections other than CMV, HCT recipients who received foscarnet for preengraftment CMV, patients who received foscarnet for < 6 days, and those with inadequate data (4 HCT and 2 SOT). The cutoff of < 6 days of therapy was selected to exclude patients in whom foscarnet was started shortly before death.

Data Collection

Data collected from the electronic medical record included: demographics and clinical variables, outcomes (death, virologic clearance), all-cause mortality, virologic data (CMV viremia vs. disease, peak viral load, antiviral resistance mutations), and treatment data, including doses and durations of antiviral therapy administered for CMV treatment prior to foscarnet therapy, and doses and durations of foscarnet. Estimated glomerular filtration rate (eGFR) at the start and end of foscarnet therapy, as well as 6 months after completion of foscarnet, was also collected.

CMV Prophylaxis and Preemptive Therapy

During this era, SOT recipients received CMV prophylaxis according to donor/recipient serostatus, as follows: donor-seropositive/recipient-seronegative (D+/R−) patients received valganciclovir 900 mg/d for 6 months (except lung recipients received ganciclovir 5 mg/kg IV Q24h for 6 months until 2013 when this was changed to ganciclovir 5 mg/kg IV q24h for 3 months followed by valganciclovir 900 mg/d for 3 months); recipient-seropositive (R+) SOT recipients received valganciclovir 450 mg/d for 3 months (except lung recipients received ganciclovir 5 mg/kg IV Q24h for 3 months until 2013 when this was changed to valganciclovir 900 mg/d for 3 months). CMV D−/R− SOT recipients received acyclovir. HCT recipients did not receive–CMV prophylaxis, but were treated with valacyclovir 500 mg BID and preemptive therapy with monitoring by CMV PCR 1 – 2 times per week from the time of transplant through Day 100, and after that, per clinician choice.

Laboratory Assays

CMV quantitative PCR assays were performed using a laboratory-developed assay based on Qiagen-artus analyte specific reagents until 4/2/2013 (copies/ml with lower limit of quantification 100 copies/ml)²⁴. From 4/2/2013 onward, an FDA-approved real-time PCR assay was performed (Cobas AmpliPrep/Cobas TaqMan CMV, results in IU/ml with lower limit of quantification 137 IU/ml). Assay verification experiments corroborated manufacturer’s copy/ml to IU/mL conversion factor; therefore, assay results prior to 4/2013 were converted from DNA copies/mL to IU/mL using a conversion factor of 0.9. Genotyping for resistance mutations was performed upon clinician request; UL97 and UL54 mutation detection were performed by direct sequencing in the Johns Hopkins Medical Microbiology Laboratory. Patients whose CMV isolates were referred for genotyping all had UL97 mutation assays, but UL54 mutation assays were performed only upon request.

Definitions

Standard definitions were used for tissue-invasive CMV disease and CMV viremia²⁵, whereby “tissue-invasive CMV disease” requires a diagnosis by tissue biopsy or immunostaining (with the exception of CMV retinitis) and “CMV viremia” refers to any detection of CMV in blood. In this study, the term “CMV viremia” was used to indicate detection of CMV in blood by PCR in patients who did not have tissue-invasive disease. Virologic clearance was defined as the achievement of 2 consecutive undetectable CMV PCR results at least 5 days apart. CMV PCR results that remained detectable below the limit of quantification were not considered to represent virologic clearance.

Ganciclovir resistance (“resistant” group) was defined as the presence of a UL97 or UL54 mutation that has been confirmed in previous studies to confer ganciclovir resistance through marker rescue²⁶. Patients who had had genotype testing performed and did not have confirmed ganciclovir resistance mutations, but who were judged by their treating clinicians to have not responded to ganciclovir or valganciclovir therapy were considered to be in the “refractory” category. Patients who did not have genotyping performed were considered to be in the “indeterminate” category as regards ganciclovir resistance. Changes in antiviral treatment were made per clinician choice without reference to a uniform definition of “refractory” (such as failure of reduction in CMV viremia after a specified period of therapy).

Renal dysfunction was defined as a >20% decrease in estimated glomerular filtration rate (eGFR) by the MDRD method. This cutoff was chosen to be consistent with most published investigations assessing change of kidney function as a metric of outcome.

Statistical Analysis

Descriptive statistics (means, medians) were performed with analyses in Stata, v. 12. Two-way comparisons of categorical variables were performed using Fisher’s exact test or chi-square test. Associations between continuous variables, such as peak viral load, and binary variables such as survival, resistant vs refractory CMV, SOT vs HCT, were tested using 1-way analysis of variance.

RESULTS

Patients

A total of 39 patients were included (22 SOT, 17 HCT, see Table 1), including 35 adult and 4 pediatric recipients. All HCT were allogeneic transplant recipients. Basic demographics and types of transplant are recorded in Table 1.

TABLE 1.

Demographic and Clinical Characteristics and Outcomes of SOT and HCT Recipients Treated With Foscarnet for Resistant/Refractory CMV

	SOT (n = 22)*	HCT (n = 17)**
M/F	16M, 6F	8M, 9F
Median Age at Transplant	54 (25–68)	33 (8–66)
CMV D+/R−	17	3
CMV D+/R+	4	4
CMV D−/R+	0	3***
CMV D−/R+	0	4
CMV D−/R−	1	0
GCV-Resistant****	13/22 (59%)	2/17 (12%)
Median Peak Viral Load*****	241 500 (765–45,500 000) IU/mL	16 100 (617–469 000) IU/mL
Median Days Post-Transplant When FOS Started	194 (93–542)	73 (29–521)
Tissue-Invasive CMV	7/22 (32%)	4/17 (24%)
Virologic Failure on FOS	6/22 (27%)	7/17 (41%)
Mortality Within 1 Year****	2/22 (9%)	10/17 (59%)
>20% Decrease in eGFR by End of FOS	12/22 (55%)	8/17 (47%)

^*SOT recipients included kidney (13), lung (4), kidney/pancreas (2), liver (2), heart (1).

^**HCT recipients included myeloablative haploidentical donor (4), myeloablative matched related donor (2), myeloablative 5/6 antigen related donor (1), myeloablative matched unrelated donor (1), nonmyeloablative haploidentical donor (6), nonmyeloablative matched related donor (1), nonmyeloablative matched unrelated donor (1), double cord transplant (1). Underlying diagnoses included acute myelogenous leukemia (8), chronic myelomonocytic leukemia/acute myelogenous leukemia (1), acute lymphoblastic leukemia (1), mixed lineage leukemia (1), dyskeratosis congenita with myelodysplastic syndrome (2), myelodysplastic syndrome (1), sickle cell disease (1), Hodgkin disease (1), and nonHodgkin’s lymphoma (1).

^***Four patients (all HCT) had missing donor information; 3 of these were R+ (CMV D−/R+)and 1 had missing recipient information also.

^****p = 0.003 for % ganciclovir-resistant in SOT vs HCT, and p = 0.001 for % mortality in SOT vs HCT; other results were not statistically significant.

^*****Units of viral load were DNA copies/mL before 4/2013 and IU/mL from 4/2013 onward: viral loads before 4/2013 were converted to IU/mL using a conversion factor of 0.9.

Foscarnet-treated patients with resistant or refractory CMV represented 0.66% of all SOT and 0.74% of all HCT performed during this time period at this center. The majority of SOT patients were in the high-risk donor-seropositive, recipient-seronegative (D+/R−) category (17/22, 77%), while 13/16 (81%) of HCT recipients were recipient-seropositive (R+). CMV serologic information was missing for 4 HCT donors and 1 HCT recipient.

Ganciclovir-Resistant and Refractory CMV Patients

UL97 resistance genotyping was performed in 32/39 patients (82%). Of these, 17 had UL97 mutations (Table 2); 14 of these were mutations that have been confirmed to confer ganciclovir resistance, while 2 had UL97 polymorphisms (H469Y, N510S) that do not confer ganciclovir resistance; these 2 were considered to be in the “refractory” category. One patient had a UL97 mutation described as ganciclovir-resistant by an outside laboratory however the specific mutation was not noted; this patient was considered in the “resistant” category. Fourteen patients had no detectable mutations. Seven patients did not undergo UL97 genotype testing, and 1 had a viral load too low to permit sequencing. These 8 patients (3 SOT, 5 HCT) were classified as “indeterminate” regarding resistance. In all, 15 patients were in the “ganciclovir-resistant” and 16 patients were in the “refractory” category, while 8 were “indeterminate”. UL54 genotype tests were performed in only 2 patients, per clinician request, neither of which revealed mutations that confer antiviral resistance. Of the 15 patients with documented ganciclovir resistance mutations, 13 were SOT and 2 were HCT; thus, ganciclovir resistance occurred more commonly in SOT than in HCT recipients (13 of 19 SOT tested vs 2 of 12 HCT tested, p = 0.009). In those with documented ganciclovir resistance, 11 of the 13 SOT recipients were D+/R− while the other 2 SOT were D+/R+; 1 of 2 HCT recipients with documented ganciclovir resistance was D−/R+ and the other was R+ with unknown donor status.

TABLE 2.

UL97 Genotype Mutations

None detected – 14

L595S – 5

A594V – 3

Both L595S and A594V – 2

C603W – 2

L595F – 1

M460I - 1

UL97 mutation, details not available – 1*

Viral load too low to amplify – 1

Test not done – 7

^*Performed at an outside laboratory

CMV Treatments and Virologic Outcomes

Tissue-invasive disease (end-organ disease) occurred in 11/39 (28.2%) patients, whereas viremia without tissue-invasive disease occurred in 28/39 (71.8%). There was no significant difference in tissue-invasive disease between the SOT and HCT groups. The median peak CMV viral load was 107,000 IU/mL for the group overall (range, 617–45,500,000). The median peak viral load for SOT recipients was 241,500 IU/mL vs 16,100 IU/mL for HCT recipients (p=NS). Two patients had 2 courses of foscarnet, 2 patients had 3 courses, and the remaining 35 each had 1 course of foscarnet. Foscarnet was administered for a median of 32 days (range, 6 – 193). Viremia cleared with foscarnet treatment in only 26/39 patients (67%), and 8/26 (31%, 5 SOT and 3 HCT) of these relapsed with viremia subsequently. The median time to 2 consecutive undetectable CMV PCR’s was 27 days (range, 17 days to over 3 years). There was no significant difference in peak viral load between patients who relapsed with viremia after foscarnet and those who did not. Of patients who did not clear with foscarnet, 6/13 (46%) had foscarnet discontinued because of toxicity. There were no statistically significant differences in virologic clearance between the ganciclovir-resistant vs. refractory groups, or between SOT and HCT recipients. Virologic clearance was not associated with survival.

In terms of other therapies, CMV hyperimmune globulin (CMVIg) was administered to 27/39 patients (69%, 15 SOT and 12 HCT) as an adjunct to antiviral therapy; 9 of these 27 had tissue-invasive disease. Cidofovir was administered to 5 patients. One patient received combination therapy with ganciclovir and foscarnet. In 7 patients who failed to achieve virologic clearance on foscarnet, but who ultimately cleared viremia, a variety of therapies were administered after discontinuation of foscarnet, including cidofovir (3), ganciclovir (3), valganciclovir (2), leflunomide (1), and CMV-specific T cell therapy (1).

Antiviral Durations and Dosing

All patients had received either ganciclovir or valganciclovir (or both) as treatment for active CMV infection prior to switching to foscarnet. In all but 2 patients (37/39 or 95%), the clinical notes cited a rising CMV viral load as a reason for the switch to foscarnet. In the remaining 2 patients (both HCT), the reasons cited were neutropenia with persistent CMV viremia in 1, and CMV viremia with worsening mental status and concomitant HHV-6 viremia in the other. In 36 patients for whom total days of ganciclovir therapy prior to foscarnet could be determined, the median was 15 days (range 3 – 88). In addition, 22/39 patients (56%) had received valganciclovir for treatment of CMV infection for a median of 44 days (range 6 – 296) prior to foscarnet, not counting days of prophylaxis. In 4 patients (all D+/R- SOT), the initial CMV viremia was breakthrough viremia while on valganciclovir prophylaxis. 10/39 patients (26%) had subtherapeutic doses of either ganciclovir or valganciclovir with respect to the manufacturer’s recommendations for dose adjustment for renal function, at some point in time prior to switching to foscarnet. Often, this occurred in the setting of delayed adjustment of dosing in the setting of improving renal function. Of these 10 patients, 7 were in the ganciclovir-resistant group and 3 were in the refractory group (p=0.2).

Foscarnet dosing information was available for all but 4 patients. Initial dosing of foscarnet in these patients was appropriate induction dosing per the manufacturer’s nomogram. Subsequently, 13 patients had doses of foscarnet decreased for changes in renal function (including 9 who cleared viremia on foscarnet and 4 who failed to clear) and 2 of these received inappropriately low doses of foscarnet at least at 1 point in time (both failed to clear viremia.) he majority of patients who underwent dose decreases still cleared viremia on foscarnet (9/13, 69%) as long as these were undertaken in accordance with standard dosing protocols.

Mortality

The 1-year all-cause mortality in this group of patients was 12/39 (31%). All deaths occurred within 4 months of the end of foscarnet therapy, and 9/12 deaths (75%) occurred within 1 month of the end of foscarnet therapy, or while still on foscarnet. Deaths were more common in HCT than in SOT recipients (p=0.001) with HCT recipients accounting for 10/12 (83%) of deaths (Table 3). Only 1 who died (8%) had a ganciclovir resistance mutation; the remaining 11 (92%) were in the refractory category (p=0.013). Tissue-invasive CMV was present in 42% of patients who died compared with 28% overall, (p = 0.48). Only 33% of patients who died had post-mortem examinations. Among the 13 patients who failed to achieve virologic clearance on foscarnet, there were 6 who died before CMV viremia cleared. The other 6 patients who died had undetectable CMV viral loads at the time of death (Table 3), and 3/6 (50%) of those had tissue-invasive disease compared with 28% overall (p = 0.35).

TABLE 3.

Characteristics of SOT and HCT Recipients Who Died After Foscarnet Salvage Therapy

Type of transplant	Virologic clearance on FOS	GCV R	Tissue- invasive disease	Autopsy?	Viral Load at Death	Peak Viral Load	Cause of death if known	Other comorbid conditions
HCT	Yes	Yes	None	No	Undet	30,100	Multiorgan failure, ARDS	MRSA BSI, Pseudomonas pneumonia
HCT	No	No	Colon	Yes	1,360	514,000	Pulmonary hemorrhage, DIC	Dyskeratosis congenita
SPK	No	No	Lung, liver	No	20,900	186,000	Respiratory failure, sepsis	VRE bacteremia
HCT	Yes	No	None	No	Undet	4,266	GVHD, GI bleeding	Liver dysfunction, HSV esophagitis
Lung	Yes	No	Lung	Yes	Undet	8,928	Diffuse alveolar damage after CMV pneumonitis	Pseudomonas pneumonia
HCT	Yes	No	None	No	Undet	164,540	ARDS, renal failure	MDS relapse, graft failure, aspergillosis
HCT	Yes	ND	Colon	No	Undet	27,200	Multiorgan failure, GVHD	Pneumonia, suspected CNS fungal
HCT	No	ND	None	No	20,672	20,672	Respiratory, multiorgan failure	Relapsed leukemia, BKV
HCT	No	No	None	No	Detected < 137	469,000	Relapsed lymphoma	CVA
HCT	Yes	No	Lung	Yes	Undet	2,080	Bronchiolitis obliterans after CMV pneumonitis	Pneumo- mediastinum
HCT	No	No	None	No	Detected < 137	8,760	Sepsis, multiorgan failure	Graft loss while on FOS
HCT	No	No	None	Yes	226	617	Sepsis, pneumonia	Encephalitis CMV vs HHV6

^*Abbreviations: GCV – ganciclovir; FOS – foscarnet; ARDS – acute respiratory distress syndrome; MRSA – methicillin-resistant Staphylococcus aureus; BSI – bloodstream infection; DIC - disseminated intravascular coagulation; SPK - simultaneous pancreas-kidney transplant; VRE – vancomycin-resistant Enterococcus; MDS – myelodysplastic syndrome; GVHD – graft-vs-host disease; GI – gastrointestinal; HSV – herpes simplex virus; CNS – central nervous system; BKV – BK virus; CVA – cerebrovascular accident; HHV-6 – human herpesvirus - 6

Renal Dysfunction and Other Adverse Effects

Renal dysfunction (eGFR decrease of >20% between the start and end of foscarnet therapy) occurred in 20/39 patients (51.2%). Eight of these 20 (40%) were kidney recipients, but renal dysfunction was not statistically more likely in kidney recipients (p=0.77). Three patients were not on hemodialysis at the start, but required renal replacement therapy by the end of foscarnet treatment; all 3 of these patients died. Two of these 3 patients had eGFR’s of >60 at the start of foscarnet therapy and 1 had an initial eGFR of 43. One patient was on hemodialysis at the start and throughout foscarnet therapy; this patient survived. Information on eGFR at 6 months after the end of foscarnet therapy was available for 25 of the 27 patients who survived to this time point; 7/25 (28%) had an eGFR that was > 20% worse than the baseline eGFR.

Supplemental IV hydration prior to each dose of foscarnet was administered to 30/31 patients for whom information was available. Electrolyte derangements were reported in 14 patients (potassium, calcium, phosphate, and magnesium). Other adverse effects attributed to foscarnet included severe nausea (6 patients), achiness and malaise (2), painful genitourinary ulcerations (1), urinary incontinence (1), edema (1), leg and abdominal cramps (1), oral ulcerations (1) and shortness of breath (1). One patient (HCT) developed neutropenia and irreversible graft loss while on foscarnet.

Outcomes in Pediatric Patients

There were 4 pediatric patients in this study; all HCT, and all had refractory CMV. Two had biopsy-proven tissue-invasive disease. Three cleared viremia on foscarnet and 1 failed to clear; 2 of 4 patients died. The patient who failed to clear viremia on foscarnet was CMV donor seronegative, recipient seropositive, and died with persistent though decreased viremia, with development of respiratory failure and pulmonary hemorrhage.

DISCUSSION

Ganciclovir-resistant CMV remains a challenging problem, particularly in (D+/R−) SOT recipients in whom the incidence of genotypically confirmed resistance is as high as 5 – 10 % of patients treated for viremia²⁶. Incidence of genotypically resistant CMV is lower in HCT (0 – 4%)²⁶; however, in both HCT and SOT recipients, a “refractory” CMV category has been recognized, in which lack of response to ganciclovir or valganciclovir occurs despite the absence of detectable resistance mutations. In the present study, almost half of the patients with documented ganciclovir resistance had prior ganciclovir/valganciclovir dosing that was below the recommended dose for the patient’s level of renal function at some point in time, so it is possible that subtherapeutic levels contributed to the development of genotypic resistance in this group. Whether suboptimal dosing is associated with “refractory” CMV without demonstrable genotypic resistance is unclear, since only 3 patients in this category had suboptimal prior dosing of ganciclovir/valganciclovir.

In the current study, both virologic responses and clinical outcomes were suboptimal in a group of transplant patients who received foscarnet. All-cause mortality in this group of patients was high, with the preponderance of deaths occurring in HCT recipients, and in the refractory (rather than ganciclovir-resistant) category. Thus, outcomes of second-line therapy for refractory CMV in highly treatment-experienced patients are poor even when genotypic resistance mutations are not confirmed. This suggests that host factors are important determinants of mortality in this patient population. It is important to recognize the existence of this “refractory” category of patients, in order to develop optimal management strategies in the future.

Although CMV-attributable mortality would be of interest, it is difficult to ascertain in this group, due to the complexity of their illnesses. One of the most common scenarios was death from multiorgan failure shortly after virologic clearance or with persistent (though decreased) CMV viral load. In such situations, it is hard to distinguish between the direct consequences of CMV replication, the indirect effects of CMV on the allograft, the toxicities of anti-CMV therapies, and other concomitant infections and comorbidities. Given that 50% (3/6) of those who died despite clearing viremia had had tissue-invasive CMV disease, the persistence of CMV in end-organs despite virologic clearance is also a possibility, although this did not reach statistical significance when compared with the 28% overall incidence of tissue-invasive disease in this cohort. In addition, only 1/3 of patients who died had autopsies performed. In most of these cases, however, the treating clinicians considered CMV to have been at least a contributing factor to the patient’s deteriorating clinical course, if not the principal cause. All deaths occurred within 4 months of the end of foscarnet therapy, and most were within 1 month of the end of foscarnet or while still on foscarnet.

Renal dysfunction occurred in over half of this group of patients, and this persisted in some patients out to 6 months after completion of foscarnet. Many other factors can contribute to renal dysfunction in these complex patients. Nonetheless, the observation of renal dysfunction occurring by the end of treatment in over 50% of patients on foscarnet is consistent with the clinical impression that renal dysfunction often occurs in patients receiving this drug, whether from foscarnet or other causes.

Only a few previous studies have reported detailed outcomes in foscarnet-treated patients with resistant/refractory CMV. A review by Bacigalupo et al¹⁶ summarizes those evaluating the use of foscarnet for prophylaxis and preemptive therapy as well as for treatment. Studies in the last 15 years that included at least 6 transplant recipients who were treated with foscarnet for established CMV infection, and that reported mortality outcomes, are summarized in Table 4. Other than the case-control study by Fisher et al²², the others are case series, and most have focused on SOT recipients. All-cause mortality rates in these studies varied, but the mortality rates reported in the studies by Pierce et al²³ (32%) and Minces et al²¹ (31%) were similar to that in the current study. Taken together, this group of studies suggest that mortality is higher in HCT recipients than SOT recipients with resistant/refractory CMV, likely reflecting host immunologic factors, underlying diseases, and comorbidities. A complicating feature is the capacity for both CMV and its treatment to cause leukopenia in HCT recipients, with consequent high risk of sepsis and other infections.

TABLE 4.

Studies Published After the Year 2000, Reporting Outcomes of 6 or More Transplant Recipients Treated with Foscarnet for Established CMV Infection

Study	Year/Center	Patients	Total n	Deaths By 1 Year	Renal Dysfunction End of FOS	Renal Dysfunction Long-Term
Current study	2015 Johns Hopkins	FOS- treated R/R SOT + HCT	39 (all FOS)	12/39 (31%)	20/39 (51%)	7/25 (24%) at 6 mos
Pierce et al21	2015 Northwestern	FOS- treated R/R SOT	31 (all FOS)	10/31 (32%)	5/21 (24%)	3/21 (14%)
Fisher et al20	2014 University of Washington	GCV-R SOT	38 cases, 110 controls	8/38 (21%)	NR	15/37 (41%) at 3 mos
Minces et al19	2014 University of Pittsburgh	GCV-R Lung Transplant	16 (14 FOS)	5/16 (31%)	10/14 (71%)	NR
Myhre et al17	2011 Oslo University	GCV-R Kidney Transplant	27 (10 FOS)	2/10 (20%)	NR	NR
Asakura et al16	2010 Nagoya University	FOS- treated HCT	65 CMV disease (all FOS)	45/65 (69%)	3%*	NR
Reddy et al15	2007 Duke University	GCV-R Lung Transplant	6 (all FOS)	1/6 (17%)	2/6 (33%)	0/6 (0%)
Isada et al2	2002 Cleveland Clinic	GCV-R SOT	13 (10 FOS)	9/10 (90%)	NR	NR**

^*Asakura et al included 320 allogeneic HCT recipients, 65 of whom received FOS for treatment of CMV disease, whereas others were treated with FOS for preemptive therapy or prophylaxis. Mortality data estimated from Fig 5B, p. 356¹⁶. Renal dysfunction data refers to the entire group which included those receiving FOS for preemptive therapy¹⁶.

^**Although overall nephrotoxicity was not reported in Isada et al, 4 of 5 (80%) of kidney transplant recipients developed graft loss².

Abbreviations: GCV-R – ganciclovir-resistant; SOT – solid organ transplant; HCT – hematopoietic cell transplant; FOS – foscarnet; R/R – resistant/refractory; NR – not reported.

Given these suboptimal results with foscarnet treatment, more information on 2other common strategies would be welcome. The use of high-dose (“supratherapeutic”) ganciclovir or valganciclovir, and the adjunctive use of intravenous immunoglobulin (IVIg) or CMV hyperimmune globulin (CMVIg), have become widespread, but data from randomized trials are lacking. Gracia-Ahufinger et al reported on 6 patients with genotypic ganciclovir resistance mutations treated with high-dose ganciclovir or valganciclovir. All patients responded to therapy, but neutropenia occurred in 50%⁵. More comparative outcomes data would be welcome, in order to identify which subgroups of patients could safely and effectively be treated with such a regimen. Regarding adjunctive intravenous immunoglobulin treatment, Florescu et al, in their international survey of CMV prevention and treatment practices in intestinal transplantation, found that 65% of programs used IVIg or CMVIg for various indications, most commonly adjunctive treatment of CMV pneumonitis, meningoencephalitis, and enteritis⁸. Although the authors commented that IVIg and CMVIg use is probably more common in intestinal transplantation than in other solid organ transplants, this still represents a substantial use of costly therapies with little evidence base outside of CMV pneumonitis in bone marrow transplant recipients⁷. A more detailed study of these practices is warranted if we are to define the optimal management for patients with resistant or refractory CMV infection.

In this study, greater mortality was observed in HCT compared to SOT recipients despite a lower prevalence of ganciclovir-resistant CMV among the former group. This observation is consistent with the report by Nichols et al of the frequent phenomenon of rising viral load in the absence of ganciclovir resistance in HCT recipients²⁷, suggesting that the biology of refractory CMV may be different between SOT and HCT recipients. All in all, a mortality range of 17 – 32% for SOT, and 59 – 69% for HCT, are probably reasonable estimates for patients treated with currently available salvage antiviral agents for resistant/refractory CMV infection.

LIMITATIONS

This study has multiple limitations, many related to retrospective design and to the information available. The group of patients reported herein was highly heterogeneous as to type of transplant, genotypic resistance, and severity of infection. CMV resistance genotype analysis was performed per clinician choice, so not all patients had this testing performed; only 2 had UL54 mutation analyses performed per specific request. In future studies patients would ideally undergo both UL97 and UL54 mutation testing in any suspected cases of resistance.

Since the decision to switch to foscarnet therapy was per clinician choice, there was no uniform definition of “refractory” CMV. In a prospective clinical trial, there would be a formal definition of “refractory”, for example a failure of the CMV viral load to achieve a prespecified fall over a prespecified period of time.

The electronic medical record at this center underwent stepwise implementation during the time period in question, and thus there was some missing information with respect to doses and durations of antiviral therapy. Different quantification tests were used to measure plasma CMV DNA during the study period, before and after 4/2013, although a conversion factor was used to convert earlier viral loads to the current IU/mL. Many of these limitations would be rectified in prospective studies using an electronic medical record with complete medication dosing information, and using the current FDA-approved CMV PCR assay as is currently the case at this center.

CONCLUSIONS

Foscarnet has been the mainstay of treatment for ganciclovir-resistant CMV. However, outcomes in transplant recipients treated with foscarnet are less than ideal, with suboptimal virologic clearance, high mortality rates, and potential contribution to renal dysfunction. In this study, genotypic resistance was more common in SOT, especially D+/R− SOT, although mortality was higher in HCT recipients. Dosing of ganciclovir and valganciclovir that was low at some point in time relative to the patient’s level of renal function may have contributed to the development of ganciclovir resistance. This study cannot provide a basis for recommendations on which patients would benefit from the switch to foscarnet, or when to contemplate such a switch. However, this series should help to provide a sense of the variety of challenges and the burden of illness in patients who are treated with foscarnet for resistant or refractory CMV. Future studies of investigational anti-CMV agents for treatment of established CMV are needed, in the hopes of achieving both greater efficacy and decreased toxicity.

ABBREVIATIONS

CMV

cytomegalovirus

SOT

solid organ transplant

HCT

hematopoietic cell transplant

CMVIg

CMV immune globulin

eGFR

estimated glomerular filtration rate

MDRD

Modification of Diet in Renal Disease Study Equation

D+/R−

donor seropositive, recipient seronegative (et seq)

GCV

ganciclovir

FOS

foscarnet