Outcomes of transplant recipients treated with cidofovir for resistant or refractory cytomegalovirus infection

Abstract

Background:

Treatment of ganciclovir-resistant (GCV-R)/refractory cytomegalovirus (CMV) infections in blood/marrow transplant (BMT) and solid organ transplant (SOT) recipients remains suboptimal. Cidofovir (CDV), a nucleotide analogue with anti-CMV activity, is nephrotoxic and oculotoxic.

Methods:

We retrospectively evaluated the outcomes of SOT and BMT patients with GCV-R/refractory CMV treated with CDV between 1/1/2008 and 12/31/2017. Data collected: baseline demographics, CMV serostatus, clinical and virologic presentations and outcomes, UL97 and UL54 genotype mutations, drug toxicities, and cause of death. Descriptive statistics were used.

Results:

16 patients received CDV for treatment of CMV: six BMT and 10 SOT. Seven (47%) of the patients had high-risk donor/recipient serostatus: six (60%) SOT were D+/R−; one (16.7%) BMT was D−/R+. Median time to CMV DNAemia was 131 days post-transplant (IQR, 37.5–230.3). Proven tissue invasive disease was present in three patients (18.8%). Twelve (75%) had genotype testing; 10 (83.3%) of those had antiviral resistance mutations. While on CDV, six (37.5%) developed nephrotoxicity, and four (25%) developed uveitis (two had both uveitis and nephrotoxicity). Eight (50%) had failure to clear CMV DNAemia despite CDV treatment. Eight (50%) of the patients died; median time to death, after initiation of CDV, was 33.5 days [IQR22–988].

Conclusions:

In the absence of good therapeutic alternatives, CDV is used in GCV-R/refractory CMV infection. However, it is associated with a substantial risk of toxicity and failure to clear CMV DNAemia, highlighting the need for development of newer and less toxic therapies. The high mortality in this group of patients underscores the severity of illness in this population.

1 |. INTRODUCTION

Ganciclovir-resistant (GCV-R) cytomegalovirus (CMV) infection is an important and growing problem among blood/marrow transplant (BMT) and solid organ transplant (SOT) patients and is associated with higher morbidity and mortality as compared to ganciclovir-sensitive CMV.^1–5 In addition to GCV-R CMV infection, there is consensus amongst experts that an entity of “refractory” CMV infection exists, described as patients who do not respond clinically and/or virologically to standard therapy with GCV, but who do not have evidence of GCV-resistance conferring mutations by genotypic analysis.^1,6 The treatment of GCV-R and refractory CMV infection remains a challenge and only a few studies have reported on the clinical and virologic outcomes of cidofovir (CDV) therapy for this purpose.^6–10

Foscarnet (FOS) and CDV are nucleotide analogues with broad antiviral activity. FOS is often considered the second-line therapy after GCV derivatives.⁸ Our group has previously published a retrospective case series of 39 transplant recipients (22 SOT and 17 BMT) who received FOS for resistant or refractory CMV. Failure to clear CMV DNAemia was noted in 33%, relapsed CMV DNAemia occurred in 31%. Renal dysfunction occurred in 51%, and mortality was 31%.¹ Though FOS is considered a second-line therapy, it is not an ideal agent to use given its renal toxicity, mortality, and virologic failure rates.

Cidofovir has activity against CMV, other human herpesviruses, and some other DNA viruses including adenovirus.^7,10 CDV is also active against many CMV strains containing UL97 mutations, in theory allowing for its off-label use in the treatment of GCV-R CMV infections. Furthermore, the pharmacokinetic profile of CDV favors once-weekly dosing.^7,9,10 However, CDV is well known for its toxicities, namely nephrotoxicity and oculotoxicity (uveitis and loss of intraocular pressure).^7–10 These toxicities make use of the drug challenging in the BMT and SOT populations with refractory or resistant CMV, some of whom have renal impairment already.^7,11 The aim of this retrospective single-center study was to evaluate the clinical and virologic outcomes of CDV treatment for GCV-R or refractory CMV in BMT and SOT patients.

2 |. MATERIALS AND METHODS

2.1 |. Patients

After obtaining approval with wavier of consent from the Johns Hopkins Institutional Review Board (IRB# 00086701), all BMT and SOT recipients who were treated with CDV for GCV-R or refractory CMV infection between 1/1/2008 and 12/31/2017 were retrospectively evaluated. Patients who had received CDV for these indications were identified through hospital pharmacy records. Patients were excluded if they received CDV as treatment for any other indication aside from CMV. Duration of follow-up was calculated from the date of the first dose of CDV. For patients who died, the last date of follow-up was the date of death, and for others, the last date the patient was seen at the transplant center.

2.2 |. Data collection

Data collected from electronic medical records included: baseline patient demographics (gender, age, type of transplant), CMV IgG serostatus of donor and recipient, details of CMV clinical and virologic course, days to CMV DNAemia from time of transplant, peak CMV DNA levels, presence of tissue-invasive CMV disease, UL97 and UL54 genotype mutations, drug toxicities (renal and ocular), virologic outcomes, and cause of death.

2.3 |. Laboratory Assays

Until April 2, 2013, CMV quantitative polymerase chain reaction (PCR) assays at this center were performed utilizing an in-house developed assay which was based upon Qiagen-artus analyte-specific reagents (copies/mL with lower limit of quantification 100 copies/mL).¹² From April 2, 2013 onward, the Cobas AmpliPrep/Cobas TaqMan CMV assay (an FDA-approved real-time PCR assay) was utilized (lU/mL with lower limit of quantification 137 lU/mL). Assay results prior to April 2013 were converted from DNA copies/mL to IU/mL using a conversion factor of 0.9 (a corroborated manufacturer’s conversion factor). All CMV qPCR assays were performed on plasma. Genotyping for UL97 and UL54 mutations was performed by direct bi-directional Sanger sequencing, on selected patients per clinician request.

2.4 |. Definitions

Cytomegalovirus end-organ disease and CMV DNAemia were defined as per CMV Drug Development Forum consensus definitions.¹³ In addition, patients who did not undergo tissue biopsies but whom clinicians treated as suspected CMV end-organ disease were reported as “suspected,” although these fell short of the criteria for “proven”, “probable,” or “possible” CMV end-organ disease in the consensus definitions.¹³ Virologic outcomes referred to clearance of CMV DNAemia vs failure to clear CMV DNAemia. Clearance of CMV DNAemia was defined as the achievement of two consecutive undetectable CMV qPCR results at least 5 days apart. Failure to clear CMV DNAemia was the inability to achieve clearance of CMV DNAemia as defined above.

Ganciclovir-resistant infection was defined as infection with an UL97 or UL54 mutation present known to confer GCV resistance.¹⁴ For the purposes of this study, refractory CMV infection was defined as failure of CMV DNA to decline by 1.0 log₁₀ during 14 days of previous therapy with other CMV antivirals prior to starting CDV, in those in whom genotype testing did not reveal antiviral resistance mutations. Those who did not have genotype testing performed were considered indeterminate. The patients in this case series were treated prior to the publication of the new definitions of refractory CMV.¹³

Nephrotoxicity was defined as an increase in serum creatinine, by the end of CDV therapy, by 1.5 times the baseline value at the start of CDV therapy. Oculotoxicity was defined as any alteration or disturbance in vision due to uveitis, iritis, or ocular hypotonia as diagnosed by an ophthalmologist.

2.5 |. Statistical Analysis

Descriptive statistics (means, medians) were performed using JMP® Pro 14.3.0. Two-way comparisons of categorical variables were performed using Fisher’s exact test or chi-square test. Associations between continuous variables, such as peak CMV DNA level, and binary variables, such as SOT vs BMT, were tested using one-way analysis of variance.

3 |. RESULTS

3.1 |. Patients

Patient demographics and CMV-related information are summarized in Table 1. A total of 16 patients were included in our analyses. Of the 16, six were BMT recipients and 10 were SOT recipients. One of the BMT recipients received a myeloablative haploidentical allogeneic transplant, four received non-myeloablative haploidentical allogeneic transplants, and one received a double cord transplant. Of the BMT recipients, two were donor-seropositive, recipient-seronegative (D+/R−), one was D−/R+, one was D+/R+, and one was D−/R−. The double cord transplant recipient’s CMV donor/recipient serostatus was unknown. With regard to the SOT patients, the majority were kidney transplant recipients (six, 60%). Amongst SOT recipients, six were D+/R−, three were D+/R+, and one patient (transplanted at another center) had missing serology data. Median time to onset of CMV DNAemia from the day of transplant in the BMT group was 37 days (range, 30.75–133.25), whereas in the SOT group this was 168.5 days (112.3–253.5). Median peak CMV DNA level was 116,850 lU/mL in the BMT group and 72,959 lU/mL in the SOT group, which was not significantly different (P = .72). Three patients (17.6%) had proven tissue-invasive CMV disease: one BMT recipient with autopsy proven CMV pneumonitis, and two (both kidney transplant recipients) with CMV colitis. In addition, there were four patients in whom CMV tissue-invasive disease was strongly suspected by clinicians and treated as such, but did not meet criteria for proven, probable, or possible CMV tissue-invasive disease (primarily because of lack of a biopsy).

TABLE 1.

Summary statistics of transplant recipients treated with CDVfor resistant/refractory CMV

	BMT/Oncology (N = 6)		SOT (N = 10)
Male	2 (33.3%)		5 (50%)
Female	4 (66.7%)		5 (50%)
Median age at transplant (y)	23 (IQR 20.8–31)		60 (IQR 43.3–60.5)
Type of Transplant	DC	1 (16.7%)	Kidney	6 (60%)
	MA HI	1 (16.7%)	Heart	2 (20%)
	NMA HI	4 (66.7%)	Lung	1 (10%)
			Liver	1 (10%)
Donor and recipient CMV IgG serostatus	D+/R−	2 (33.3%)	D+/R−	6 (60%)
	D−/R+	1 (16.7%)	D−/R+	0
	D+/R+	1 (16.7%)	D+/R+	3 (30%)
	D−/R−	1 (16.7%)	D−/R−	0
	D?/R?	1 (16.7%)	D?/R?	1 (10%)
Median time to CMV DNAemia from transplant (d)	37 (IQR 30.8–133.3)		168 (IQR 112.2–253.5)
Median peak CMV viral load, IU/mLa	116 850 (IQR 16,143.8–2 582 500)		72 959 (IQR 8694.3–759 750)
Tissue-invasive CMV diseaseb	3 (50%)		4 (40%)

Abbreviations: CMV, cytomegalovirus; DC, double cord; MA HI, myeloablative haploidentical; NMA HI, non-myeloablative haploidentical; PCR, polymerase chain reaction.

^aCMV viral load as detected by PCR assay was reported in DNA copies/mL before 4/2013, and lU/mL from 4/2013 onward. Conversion factor of 0.9 was used to convert from copies/mL to lU/mL.

^bThis includes biopsy proven tissue-invasive disease (one pneumonitis, two colitis) and clinically suspected cases but without biopsies performed (one colitis, one colitis/pneumonitis, one pneumonitis, and one meningitis/hemorrhagic cystitis/pneumonitis).

3.2 |. GCV-R and refractory CMV patients

Genotyping was performed in 12 (eight SOT and four BMT) of the 16 patients (Table 2). Nine of 12 patients tested (75%) had genotypic mutations, of whom eight had UL97 mutations known to confer GCV resistance; of these eight, two had more than one UL97 mutation present simultaneously. One patient had only a UL54 mutation (A809 A/V) and no UL97 mutations. Three did not have any genotypic mutations detected; these were designated as having refractory CMV. Those in whom resistance testing was not performed (four patients) were designated as indeterminate with respect to resistant or refractory CMV. Of the 12 with genotyping results, five had UL54 testing done. Of these five, three had UL54 mutations present. One of these had a UL54 mutation conferring both GCV and FOS resistance. The other two patients had mutations which conferred FOS resistance and CDV resistance, respectively.

TABLE 2.

CMV resistance: UL97 and UL54 genotype mutations

CMV genotype test result	Number of cases with result	Resistance phenotype
UL97a
A594V	3	GCV-resistance
L595S	3	GCV-resistance
C603W	2	GCV-resistance
A594A/P	1	GCV-resistance
E596G/E	1	GCV-resistance
UL54b
A692S	1	FOS-resistance
A809A/V	1	GCV + FOS-resistance
G841G/A	1	CDV-resistance
Not detected	3	–
Test not done	4	–

Note: 12 total patients had UL97 genotype evaluations: 10/13 (8 SOT and 2 BMT) had a UL97 mutation conferring GCV resistance.

Abbreviations: FOS, foscarnet; GCV, ganciclovir.

^aTwo patients had more than one UL97 mutation present (please refer to Table 4).

^bThree patients had UL97 and UL54 mutations present simultaneously. One patient only had a UL54 mutation present (please refer to Table 4).

3.3 |. CMV treatment

All patients had received intravenous GCV and/or oral valganciclovir (VGCV) prior to receiving CDV. Nine patients (five in the BMT group and four in the SOT group) had also received FOS prior to receiving CDV. Reasons for choosing CDV, as reflected in clinical notes, were FOS toxicity or failure (7), FOS national shortage (1), GCV toxicity or failure and desire to avoid FOS (6), treatment of CMV + BKV (1), and treatment of CMV + adenovirus (1).

The median time to starting CDV after the first positive CMV PCRwas 90 days in the BMT group, and 115.5 days in the SOT group. The BMT group and the SOT group received a median of 30 days and 16 days, respectively, of CDV therapy (range, 8–33 days between the first and last doses of CDV for the combined group). Both BMT and SOT groups received a median of two doses of CDV. Dosing schedules of CDV are shown in Table 3. Cidofovir dosing was per clinician preference and was either dosed weekly (12 patients) or weekly × 2 doses followed by every 2-week doses (four patients). Concomitant intravenous hydration and probenecid (typically 2 g at 3 hours prior to, and 1 g at 2 hours and 8 hours after the infusion) were administered. The initial dose of CDV was typically 5 mg/kg (or 3 mg/kg for renal dysfunction); the most common maintenance dosing was 5 mg/kg for the every 2-week schedule and 3 mg/kg for the weekly schedule (adjusted for renal dysfunction). Four in the BMT group and three in the SOT group also received CMV immune globulin (CMV Ig, Table 3). Two patients had received nelfinavir (NFV), in an off-label use based on reports that NFV has broad activity in vitro against a range of herpesviruses.¹⁵

TABLE 3.

Treatment of CMV, and outcomes of treatment

	BMT/Oncology (N = 6)	SOT (N = 10)	Total (N = 16)
Treated with GCV/VGCV before CDV	6 (100%)	10 (100%)	16 (100%)
Treated with FOS before CDV	5 (83.3%)	4 (40%)	9 (56.3%)
Median time to CDV after first CMV+ (d)	90 (IQR 43–230.75)	112 (IQR 21–154)	112 (IQR 38–152)
Median duration CDV received (d)	30 (IQR 15.25–68.25)	16 (IQR 8–35.5)	21.5 (IQR 8.3–47.3)
Median number of CDV doses received	3 (IQR 2–10)	2 (IQR 1–4)	3 (IQR 1–4)
CDV dosing schedule weekly × 2 doses then every 2 wka	1	3	4
CDV dosing schedule weeklya	5	7	12
CMV Immune globulin received	5 (83.3%)	3 (30%)	8 (50%)
GCV/VGCV given after CDV therapy	2 (33.3%)	3 (30%)	5 (31.3%)
Uveitis	1 (16.7%)	3 (30%)	4 (25%)
Nephrotoxicityb	3 (50%)	3 (30%)	6 (37.5%)
Recovery of renal functionc	0	1 (10%)	1 (6.3%)
Failure to clear CMV DNAemiad	4 (66.7%)	4 (40%)	8 (50%)
Deathe	4 (66.7%)	4 (40%)	8 (50%)
Median time to death (days) for patients who died (4 BMT, 4 SOT)	667.5 (range, 13–2606)	28.5 (range 21–53)	33.5 (IQR 22–988)

Abbreviations: CDV, cidofovir; FOS, foscarnet; GCV/VGCV, GCV, ganciclovir; VGCV, valganciclovir.

^aListed as weekly if only two doses were received, 1 wk apart.

^bA 1.5× rise in creatinine from start to end of CDV therapy.

^cRecovery to eGFR at time CDV was started, or improved eGFR at 6 months after stopping CDV therapy.

^dFailure to clear DNAemia was defined as not having undetectable CMV DNA level on two consecutive evaluations that occur 5 or more days apart.

^eTime to death in days from date of initiation of CDV.

3.4 |. CMV DNAemia outcomes

Clearance of DNAemia occurred in eight (50%) patients (Table 3) at a median time of 21 days from the start of CDV. Failure to clear DNAemia during CDV therapy was observed in the other eight patients: four in the BMT group and four in the SOT group. Failure to clear DNAemia occurred in four of eight patients with UL97 GCV-R mutations, and in three of four who did not have genotyping performed. All three with UL54 mutations present, had failure to clear DNAemia (Table 4). The median peak CMV DNA level at any time during the CMV episode was 0.7 log higher in those who failed to clear CMV DNAemia on CDV as compared with those who cleared (111 515 lU/mL vs 25 525 lU/mL), but this did not reach statistical significance. For the 8/16 (50%) of patients who achieved a 1-log decrease in CMV viral load on CDV, the median time to a 1-log decrease was 12 days from the start of CDV (IQR, 8–18.5).

TABLE 4.

Summary of each case evaluated

Patient Diagnosisa, age (y), gender	CMV serostatusb	Antiviral historyc	CMV End-organ diseased	Transplant to CDV initiation (d)e	Peak CMV viral load (lU/mL)	Genotype	Duration of CDV (d)	CMV DNAemia outcomeh	Toxicity	Clinical outcome	Presumed cause of death
OHT, 38, M	D+/R−	GCV, VGCV, FOS	Not present	468	107 030	L595Sf	18	Failure	Nephrotoxicity	Alive	–
KTx, 67, F	D+/R−	GCV, FOS	Proven colitis	292	2 097 000	L595Sf	84	Cleared	Uveitis	Deceased	CVA
BMT, 21, F	D?/R?	GCV, FOS	Not present	96	18 605	Not done	19	Failure	None	Deceased	Disease progression
KTx, 59, F	D+/R−	VGCV, GCV	Suspected colitis	292	38 888	L595Sf	6	Cleared	None	Alive	–
KTx, 30, F	D+/R+	VGCV, GCV	Suspected colitis, pneumonitis	392	18 242 969	Negative for mutations	12	Cleared	None	Alive	–
LTx, 64, M	D?/R?	VGCV, GCV	Not present	278	624	Not done	12	Cleared	None	Deceased	Sepsis, NOS
KTx, 52, M	D+/R+	VGCV, GCV	Not present	49	12 162	A594Vf	29	Cleared	None	Alive	–
LgTx, 65, M	D+/R−	VGCV, GCV	Not present	203	7804	A594Vf	27	Cleared	Uveitis	Alive	–
KTx, 70, F	D+/R+	GCV	Proven colitis	267	8991	Not done	21	Failure	None	Deceased	SDH, CVA
BMT, 20, F	D−/R+	GCV, FOS	Suspected pneumonitis	61	8760	Negative for mutations	7	Cleared	None	Deceased	Disease progression
BMT, 24, F	D+/R−	VGCV, GCV, FOS	Not present	518	196 000	Negative for mutations	83	Cleared	Uveitis + Nephrotoxicity	Alive	–
OHT, 45, F	D+/R−	GCV, FOS, NFV	Not present	236	116 000	E596G/Ef, A594A/Pf, A692Sg	51	Failure	Uveitis + Nephrotoxicity	Alive
BMT, 22, M	D+/R+	VGCV	Not present	446	6 100 000	Not done	33	Failure	Nephrotoxicity	Alive	–
KTx, 62, M	D+/R−	GCV, FOS, LEF, NFV	Not present	402	314 000	A594Vf, C603Wf	13	Failure	Nephrotoxicity	Alive	–
BMT, 26, F	D+/R−	GCV, FOS	Suspected HCd, meningitis, and pneumonitis	190	1410 000	A809A/Vg	53	Failure	Nephrotoxicity	Deceased	Disease progression
BMT, 46, M	D−/R−	GCV, FOS	Proven pneumonitis	142	37 700	C603Wf, G841G/Ag	26	Failure	None	Deceased	DAD due to CMV pneumonitis

Abbreviations: CVA, cerebral vascular accident; DAD, diffuse alveolar damage; disease progression, progression of underlying malignancy; SDH, subdural hematoma.

^aBMT vs SOT. SOT categories: OHT, orthotopic heart transplant; KTx, kidney transplant; LTx, liver transplant; LgTx, lung transplant.

^bCMV donor/recipient serostatus.

^cAntiviral history prior to initiation of cidofovir (CDV): VGCV, valganciclovir; GCV, ganciclovir; FOS, foscarnet; NFV, nelfinavir; LEF, leflunomide.

^dAddresses if end-organ, tissue invasive disease is present or not, and if present, what organ is involved. HC, hemorrhagic cystitis.

^eDays between transplant date and initiation of CDV.

^fUL97 genotype (see Table 2 for resistance conferred by specific mutation).

^gUL54 genotype (see Table 2 for resistance conferred by specific mutation).

^hVirologic outcome: Cleared vs failure. Cleared is achieving two undetectable CMV DNA levels at least 5 days apart. Failure is not achieving clearance.

3.5 |. Toxicities and mortality

Six patients developed nephrotoxicity (37.5%) and four developed ocular toxicity (uveitis; 25%); two patients experienced both nephrotoxicity and uveitis. Thus, in total, eight patients experienced either nephrotoxicity, uveitis, or both (Table 4). The median number of doses of CDV received in those who acquired nephrotoxicity was three (range, two-seven doses); and in those who acquired uveitis was five (range, four-seven doses). Those who did not experience any toxicities received a median of one dose of CDV. Although this was not statistically significant when compared to those who experienced any toxicity (P = .55), it is clinically meaningful. Of note, 3/6 (50%) of those who had experienced nephrotoxicity had an eGFR of less than 60 mL/min prior to CDV treatment, and 5/6 of those with nephrotoxicity had had prior FOS therapy. Of nine patients who had received FOS prior to starting CDV, 5/9 (56%) developed nephrotoxicity on CDV, whereas only 1/7 (14%) of those without prior FOS therapy developed nephrotoxicity (P = .14). In the six patients who experienced nephrotoxicity, the median creatinine rise was 1.3 mg/dL between the creatinine values at the start and end of CDV (IQR, 1.0–2.0). Two patients required new renal replacement therapy, and both had received FOS before CDV.

Eight deaths occurred (4 BMT, 4 SOT); the median time to death from the start of CDV therapy was approximately 30 days in the BMT and SOT groups (Table 3). Three of eight patients (37.5%) who failed to clear CMV DNAemia with CDV died. One death was attributed by the treating clinicians to CMV, as diffuse alveolar damage attributed to CMV pneumonitis was seen on autopsy. Of the other seven patients who died, three deaths were attributed to progression of underlying malignancy, and four to other illnesses (sepsis, cerebrovascular accident, subdural hematoma with cerebrovascular accident, and chronic lung allograft dysfunction seven years after the CMV episode).

The median time from initiation of CDV therapy to death, for the eight patients who died, was 33.5 days (range, 13–2606, IQR, 22–988); median time to death for SOT was 667.5 days and for BMT was 28.5 days (P = .1725). All but two of these patients (both SOT) died within 2 months of starting CDV. By contrast, the median duration of follow-up for patients who survived, from initiation of CDV therapy to the date of their most recent visit to the transplant center, was 2611.5 days (range, 669–3494, IQR 1441–3377).

4 |. DISCUSSION

This study summarizes the clinical and virologic outcomes, toxicities, and mortality of a group of 16 patients (10 SOT, six BMT), who were treated with CDV for resistant or refractory CMV infection (Table 4). These outcomes are problematic and highlight the need for new therapies for these patients.

Available treatment options for CMV infection other than GCV and VGCV are limited to FOS and CDV. Foscarnet is generally considered the next-line therapy after GCV derivatives.¹⁵ Our group and others have reported suboptimal outcomes with FOS for CMV, with respect to virologic failure, renal dysfunction, and overall mortality.¹ Cidofovir is somewhat less frequently used than FOS for therapy of GCV-R and refractory CMV, but also carries a considerable risk of toxicities, in the form of nephrotoxicity and oculotoxicity.

Information on CDV for CMV treatment is more limited than that for FOS. One larger series from Ljungman et al, on behalf of the Infectious Diseases Working Party of the European Group for Blood and Marrow Transplantation,¹⁰ described 82 BMT recipients who received CDV for CMV (20 for CMV disease, and the others for pre-emptive therapy). One-half of 20 patients treated for CMV disease responded, which was similar to our cohort.¹⁰ Nephrotoxicity lasting beyond CDV therapy occurred in 25%¹⁰; only slightly lower than our study (37.5%). In the pediatric BMT population, Cesaro et al reported responses to CDV in five of eight patients with CMV antigenemia.¹¹ In the SOT population, Bonatti et al,⁸ reported on a group of nine SOT recipients with CMV treated with CDV, either for resistant/refractory CMV or for CMV with GCV-associated neutropenia. In that series, seven of nine achieved clearance of DNAemia and nephrotoxicity was common.⁸

In our case series of 16 patients, 37.5% developed nephrotoxicity and 25% developed oculotoxicity. All but one of those who developed nephrotoxicity while on CDV had been treated previously with FOS, potentially underscoring the cumulative effects of sequential courses of nephrotoxic drugs. Furthermore, our study highlights the risk of failure to clear CMV DNAemia associated with CDV, as 50% of the patients failed to achieve undetectable CMV DNA. This failure to clear CMV DNAemia occurred despite the long half-life of phosphorylated CDV metabolites such as cidofovir diphosphate (> 24 hours).¹⁶ Even in patients who cleared CMV DNAemia, this clearance is not necessarily entirely attributable to CDV, since the host immune response, and modulation of immunosuppression, may influence the likelihood of DNAemia clearance.¹⁷

All-cause mortality in our cohort of CDV-treated patients was high at 50% overall (67% of BMT, and 40% of SOT recipients). Death occurred a median of 33.5 days after initiation of CDV therapy, underscoring the severity of illness in this group of patients. It is often difficult to determine if mortality is or is not attributable to CMV; in this case series, only one patient’s death was attributed to CMV on the basis of autopsy findings. However, it is also possible that the immunomodulatory effects of CMV, and/or the toxicities of anti-CMV therapies, may have contributed indirectly to mortality in these patients. The short survival time in those who died also meant that most of these patients did not undergo UL54 genotype analysis (to assess for mutations conferring resistance to CDV).

Our retrospective study underscores the risks associated with CDV therapy, failure of CDV therapy, and the mortality associated with GCV-R and refractory CMV infection. This information emphasizes the limitation of CMV and may serve as a stimulus to the development of more effective and less toxic CMV drugs that have different mechanisms of action, as well as to guide further discussions on combination therapy for treatment of GCV-R CMV, such as the use of FOS and letermovir (a recently licensed drug for CMV that inhibits CMV replication by binding to the terminase complex, pUL51, pUL56 and pUL89)^18–20 or CDV and letermovir.²¹ Other combination regimens to also consider include maribavir (an inhibitor of CMV UL97 kinase) with FOS, CDV, or letermovir,²² and artemisinin derivatives along with GCV, CDV, letermovir, or maribavir.^23,24

4.1 |. Limitations

This study had various limitations inherent to the nature of a retrospective study design with small numbers of patients. The patients were a heterogeneous group of SOT and BMT patients with varying degrees of severity of CMV infection, and resistant or refractory CMV status. While UL97 genotyping data were available for 12 of 16 patients, only five had UL54 genotyping performed (primarily due to the short survival time of those who failed CDV). Recent results have highlighted the difficulties of standardizing plasma viral load results, which could affect adjudication of clearance of DNAemia, particularly with a change of CMV DNA assay in 2013.²⁵ Lastly, the electronic medical record systems at our institution underwent changes during the study period, which led to some missing data and difficulties in determining exact length of CDV therapy.

5 |. CONCLUSIONS

Cidofovir is one of the few options currently available for GCV-resistant or refractory CMV infection. As this study demonstrates, its use is associated with substantial renal and ocular toxicity and suboptimal virologic outcomes. Less toxic and more effective antiviral options that target novel viral proteins or enzymes are urgently needed for patients with such infections.

Abbreviations:

BMT

blood or marrow transplant

CDV

cidofovir

CMV

cytomegalovirus

FOS

foscarnet

GCV

ganciclovir

GCV-R

ganciclovir-resistant

PCR

polymerase chain reaction

SOT

solid organ transplant

VGCV

valganciclovir