Clinical events and healthcare resource utilization associated with neutropenia and leukopenia among adult kidney transplant recipients receiving valganciclovir

Abstract

BACKGROUND

Cytomegalovirus (CMV) prophylaxis with valganciclovir and ganciclovir is associated with elevated neutropenia and leukopenia risk in kidney transplant recipients, although the impact of these events on healthcare resource utilization (HCRU) and clinical outcomes is unclear.

AIM

To quantify clinical events and HCRU associated with neutropenia and leukopenia among adults receiving valganciclovir and/or ganciclovir post-kidney transplantation.

METHODS

Adult kidney transplant recipients receiving valganciclovir and/or ganciclovir prophylaxis were identified in the TriNetX database from 2012 to 2021. Patient characteristics were evaluated in the 1-year period pre-first transplant. HCRU and adjusted event rates per person-year were evaluated in follow-up year 1 and years 2-5 after first kidney transplantation among cohorts with vs without neutropenia and/or leukopenia.

RESULTS

Of 15398 identified patients, the average age was 52.39 years and 58.70% were male. Patients with neutropenia and/or leukopenia had greater risk of clinical events for CMV-related events, opportunistic infections, use of granulocyte colony stimulating factor, and hospitalizations (relative risk > 1 in year 1 and years 2-5). Patients with vs without neutropenia and/or leukopenia had higher HCRU in year 1 and years 2-5 post kidney transplantation, including the mean number of inpatient admissions (year 1: 3.47 vs 2.76; years 2-5: 2.70 vs 2.29) and outpatient visits (48.97 vs 34.42; 31.73 vs 15.59, respectively), as well as the mean number of labs (1654.55 vs 1182.27; 622.37 vs 327.89).

CONCLUSION

Adults receiving valganciclovir and/or ganciclovir prophylaxis post-kidney transplantation had greater risk of neutropenia and/or leukopenia, which were associated with higher clinical event rates and HCRU up to 5 years post-transplantation. These findings suggest the need for alternative prophylaxis options with lower myelosuppressive effects to improve patient outcomes.

Key Words: Clinical outcome; Healthcare resource use; Kidney transplant; Leukopenia; Neutropenia; Ganciclovir; Valganciclovir

Core Tip: This observational study examined the impact of neutropenia and leukopenia up to 5 years in kidney transplant recipients who received valganciclovir or ganciclovir as cytomegalovirus (CMV) prophylaxis. The results showed that neutropenia and/or leukopenia were associated with increased risks of CMV-related events, opportunistic infections, hospitalizations, and use of granulocyte colony-stimulating factor. Additionally, patients with neutropenia and/or leukopenia had higher healthcare resource utilization, including more inpatient admissions, outpatient visits, and labs. These findings highlight the need for alternative prophylactic treatments for CMV with fewer myelosuppressive effects to improve outcomes in kidney transplant recipients.

INTRODUCTION

Cytomegalovirus (CMV) is a widespread herpesvirus with a global seroprevalence of approximately 80% in the adult population[1]. CMV infection is typically asymptomatic and self-limited in immunocompetent individuals, becoming latent following primary infection[2]. However, CMV infection imposes an increased risk of morbidity and mortality in patients with compromised immune function, including kidney transplant recipients[3]. According to the Organ Procurement & Transplantation Network, an estimated 23315 kidney transplants were performed in the United States in 2023, comprising approximately half of all transplants[4]. In this patient population, CMV serostatus of the kidney donor and recipient is the primary risk factor for CMV disease, which is defined as infection in tissue accompanied by other clinical symptoms like fever, leukopenia, arthralgia, or thrombocytopenia[5,6]. Approximately 20% of kidney transplant recipients receive an organ from a CMV seropositive donor[6-8] and seropositivity has been increasing over time in the United States, with further increases predicted[9].

CMV DNAemia and disease can result in life-threatening complications (i.e., hepatitis, nephritis, meningitis, encephalitis, etc.), allograft rejection, and death in kidney transplant recipients[10]. A 2022 systematic review reported that the rates of CMV-related mortality were up to 5.3% among kidney and/or liver transplant recipients[11]. Therefore, CMV prevention strategies are crucial to avoid reactivation of latent infection or reinfection with a novel strain post-transplant[6,12]. Prior to the introduction of letermovir, the previous standard of care for CMV prophylaxis was antiviral therapy with valganciclovir or, less commonly, ganciclovir[12]. The use of these therapies 3 to 6 months post-kidney transplant is shown to reduce the risk of CMV infection[6,13]. However, leukopenia and neutropenia are common adverse events associated with valganciclovir and ganciclovir due to the drugs’ myelosuppressive activity[14-17], ranging from 11% to 37% among treated kidney transplant recipients[15,18-20]. Neutropenia and leukopenia can lead to interruption of CMV prophylaxis, changes in the dosing of immunosuppressants, or use of granulocyte colony stimulating factor (G-CSF)[21,22], which can impact CMV infection, opportunistic infections, transplant success, or survival outcomes[21,23]. Additionally, myelosuppressive events can impose substantial economic burden on kidney transplant recipients and healthcare systems due to increased rates of hospitalizations or higher healthcare costs, particularly among patients at already high risk[24,25]. Thus, there is a need for risk assessment among kidney transplant recipients to identify patients most likely to benefit from CMV prophylaxis while minimizing the risk of myelosuppression.

The burden of neutropenia and leukopenia for United States kidney transplant recipients has not been established under the current standard of care for CMV prophylaxis or quantified using a large and recent dataset for this patient population. Thus, the objective of this study was to estimate rates of key clinical events and evaluate the healthcare resource utilization (HCRU) in year 1 and years 2-5 following kidney transplantation among transplant recipients who did or did not develop neutropenia and/or leukopenia after being treated with valganciclovir and/or ganciclovir.

MATERIALS AND METHODS

Data source

This retrospective, observational, nationwide cohort study of adult kidney transplant recipients in the United States used electronic medical records (EMRs) from the TriNetX database. TriNetX is a global federated health research network with real-time updates of anonymized EMRs, representing over 300 million patients. TriNetX ensures data accuracy through rigorous harmonization, cleaning, and quality checks, mapping data to standardized clinical terminologies. TriNetX datasets include anonymized patient data that is pseudonymized and/or de-identified, including demographics, diagnoses, procedures, labs and medications, HCRU, and mortality.

No identifiable protected health information or data were accessed or collected during the study; thus, ethical review was not required.

Study design

Adults who had undergone their first kidney transplant procedures between January 1, 2012 to October 31, 2021 were identified in the database. The date of the first observed kidney transplant procedure was defined as the index date. Patient demographics and clinical characteristics were evaluated during the year prior to the first kidney transplant, defined as the baseline period. Patients were followed until the end of the last encounter record for up to five years after the index date (i.e., follow-up period). The rates of clinical events and HCRU were evaluated during the follow-up period.

Population

Patients were eligible for inclusion if they: (1) Had undergone their first kidney transplant in the United States during January 1, 2012 and October 31, 2021; (2) were aged 18 years or older at the index date; (3) had continuous observable care during the follow-up period, defined as ≥ 1 encounter record more than 365 days post-index unless the patient had died within that time; and in addition; and (4) had continuous observable care pre-index, defined as ≥ 1 encounter record more than 365 days before the index date. Patients were further required to have ≥ 1 non-missing laboratory test result (irrespective of the findings of the lab value) each for absolute neutrophil count (ANC) and white blood cell count (WBC) within 365 days post-index. Due to the nature of retrospective EMR data, not all lab values were fully recorded. Patients without documented lab values for ANC or WBC needed to evaluate the occurrence of neutropenia or leukopenia within 365 days post-index were therefore excluded. Finally, patients were required to have ≥ 1 order, prescription, or administration of valganciclovir or ganciclovir with a start date on or within 30 days after the index date.

To evaluate both the short-term and long-term impact of neutropenia and/or leukopenia, four patient cohorts were created: (1) Patients with neutropenia and/or leukopenia during year 1 of follow-up; (2) Patients with neutropenia and/or leukopenia during years 2-5 of follow-up; (3) Patients without neutropenia and leukopenia during year 1 of follow-up; and (4) Patients without neutropenia and leukopenia during years 2-5 of follow-up. The cohort with neutropenia and/or leukopenia in year 1 and years 2-5 of follow-up was independently defined and included patients who had a positive result for neutropenia (ANC < 1000/μL) and/or leukopenia (WBC < 3500/μL) during the respective timeframe[13,14]. The cohorts without neutropenia and leukopenia in year 1 and years 2-5 of follow-up were also independently defined and included patients who did not have a positive result for neutropenia and leukopenia but had a non-missing lab value during the respective timeframes.

Outcomes

All outcome measures were evaluated for the four patient cohorts, respectively.

Clinical outcomes: Clinical outcomes included recurring events (i.e., CMV DNAemia with a lower limit of quantification of 137 IU/mL, CMV disease based on diagnosis, opportunistic infections excluding CMV based on diagnosis, G-CSF use based on prescriptions, and all-cause rehospitalization); one-time events (i.e., acute graft rejection, graft loss, and all-cause mortality); and the proportion of patients with no non-mortality clinical events.

Adjusted clinical event rates per person-year for the recurring and one-time events were evaluated for each of the patient cohorts, after adjusting relevant baseline characteristics. The proportion of patients with no non-mortality clinical events was calculated as proportion of patients with no clinical events defined as above except for mortality.

HCRU: HCRU outcomes included inpatient admissions, emergency department (ED), outpatient, other, and lab visits after the index kidney transplantation hospitalization discharge date. Proportion of patients with each type of visit and the number of visits among those with at least one visit were assessed.

A patient observation approach was used for the HCRU assessment by year by cohort. In each year’s calculation, patients with a full year of continuous care were included. In the pooled HCRU outcomes in years 2-5, a single patient could contribute to each of years 2-5 as long as this patient had a full year of observable follow-up in each period.

Covariates

Demographic characteristics reported during the baseline period for the overall patient population included sex, age at index, and United States region, etc.

Clinical characteristics reported during the baseline period included use of induction immunosuppressive therapy (7 days before to 15 days after index date), initial maintenance immunosuppressive therapy (within 3-month window post-index), calendar year of transplant, Charlson Comorbidity Index (CCI) comorbidities, and modified CCI score excluding renal disease.

Baseline characteristics not aforementioned were excluded from the covariates for one or more of the following reasons: (1) For binary variables, > 90% or < 10% prevalence and (2) Collinearity with other variables included in model.

Statistical analyses

Patient characteristics were described for the overall patient population. Means ± SD were reported for continuous variables; frequency counts and percentages were reported for categorical variables.

For recurring clinical events, adjusted clinical event rates were assessed using negative binomial models as count variables and calculated as the sum of the predicted number of events divided by the sum of person years in patient follow-up. Adjustment of baseline characteristics included the initial list of covariates: Age, sex, region, induction immunosuppressive therapies (corticosteroid therapy, thymoglobulin/anti-thymocyte globulin, basiliximab, alemtuzumab), initial maintenance therapies (corticosteroid therapy, mycophenolate), calendar year at index date, and modified CCI score. Covariates with P < 0.05 were included in the final model. The SE of the adjusted event rates were calculated as the SE of the mean adjusted clinical events per patient year.

For one-time events, adjusted clinical event rates were assessed using adjusted Kaplan-Meier (KM) curves as time-to-event variables. The initial list of covariates with P < 0.05 in Cox regressions were included in the adjustment of KM curves. Parametric distributions among exponential, Weibull, log-normal, log-logistic, and generalized Gamma of the adjusted KM curves were determined based on the lowest Akaike information criterion and Bayesian information criterion values. The adjusted clinical event rate at year 1 was the adjusted event rate from KM at year 1; the adjusted clinical event rate in years 2-5 was calculated as the adjusted event rate from KM at year 5 minus the adjusted event rate from KM at year 1. The SE of the adjusted event rates was obtained by the KM curves in year 1 and was assumed to be the same in years 2-5. All analyses were performed using SAS^® Studio release 9.04 or R Version 4.0.3.

RESULTS

Sample selection

A total of 15398 adults who had undergone their first kidney transplantation and who subsequently received valganciclovir and/or ganciclovir prophylaxis within 30 days of transplantation were included in the study population (Figure 1).

Figure 1 Sample selection.

Baseline characteristics

Patient demographic and clinical characteristics are reported in Table 1. The average age of patients included in the study was 52.39 (SD: 13.18) years, approximately 60% were male, and half (51.24%) were located in the South. The most common initial induction therapy was thymoglobulin/anti-thymocyte globulin (36.17%). Almost all patients used calcineurin inhibitors as initial maintenance immunosuppressive therapy (99.31%), primarily tacrolimus (99.04%). Almost all patients used a corticosteroid therapy within 3 months post kidney transplantation (91.93%). The average modified CCI score was 1.22 (SD: 1.97) and over half (52.37%) of patients had a score of 0. The most common CCI conditions were chronic diabetes (26.74%), congestive heart failure (12.67%) and peripheral vascular disease (12.82%).

Table 1 Patient baseline characteristics.

Characteristics	n = 15398
Age (years), mean (SD)	52.39 (13.18)
Age category, n (%)
18 to < 45	4395 (28.54)
45 to < 65	8111 (52.68)
65 to < 75	2670 (17.34)
≥ 75	222 (1.44)
Sex, n (%)
Male	9039 (58.70)
Female	6359 (41.30)
Region1, n (%)
South	7890 (51.24)
Northeast	4324 (28.08)
Midwest	1951 (12.67)
West	1233 (8.01)
Induction immunosuppressive therapy, n (%)
Thymoglobulin/anti-thymocyte globulin	5570 (36.17)
Basiliximab	2236 (14.52)
Alemtuzumab	1524 (9.90)
Rituximab	143 (0.93)
Initial maintenance immunosuppressive therapy, n (%)
Calcineurin inhibitors	15292 (99.31)
Tacrolimus	15250 (99.04)
Cyclosporine non-modified	451 (2.93)
Cyclosporine modified	0 (0.00)
Mycophenolic acid	12675 (82.32)
Azathioprine	330 (2.14)
Sirolimus	165 (1.07)
Everolimus	133 (0.86)
Corticosteroid therapy within 3 months post kidney transplantation, n (%)	14156 (91.93)
Prednisone	14048 (91.23)
Methyl prednisone	2927 (19.01)
Prednisolone	237 (1.54)
Calendar year of transplant, n (%)
2012	780 (5.07)
2013	1063 (6.90)
2014	1158 (7.52)
2015	1564 (10.16)
2016	1791 (11.63)
2017	2087 (13.55)
2018	2378 (15.44)
2019	2638 (17.13)
2020	1933 (12.55)
2021	6 (0.04)
Modified CCI2, mean (SD)	1.22 (1.97)
Modified CCI category, n (%)
0	8064 (52.37)
1-2	4669 (30.32)
3-4	1661 (10.79)
≥ 5	1004 (6.52)
CCI comorbidity, n (%)
Chronic diabetes	4118 (26.74)
Congestive heart failure	1951 (12.67)
Peripheral vascular disease	1974 (12.82)
Chronic pulmonary disease	1351 (8.77)
Mild to moderate diabetes	1112 (7.22)
Mild liver disease	1006 (6.53)
Cerebrovascular disease	1000 (6.49)
Myocardial infarction	732 (4.75)
Cancer	704 (4.57)
Rheumatologic disease	623 (4.05)
Moderate liver disease	516 (3.35)
Metastatic solid tumor	308 (2.00)
AIDS	298 (1.94)
Peptic ulcer disease	205 (1.33)
Hemiplegia	79 (0.51)
Dementia	27 (0.18)

¹Region was categorized based on United States Census Bureau.

²Charlson Comorbidity Index (CCI) was calculated based on a modification of the method described in Quan et al[28] (2005).The Quan method was modified to omit renal disease from the calculation of CCI and from the list of comorbid conditions included.

AIDS: Acquired immunodeficiency syndrome; CCI: Charlson Comorbidity Index.

Clinical events

Recurring events: The rates of recurring clinical events are reported in Table 2. In year 1, 11524 (74.84%) patients with neutropenia and/or leukopenia and 3874 (25.16%) without neutropenia and leukopenia were included in the analysis; in years 2-5, 5583 (38.19%) and 9035 (61.81%) patients, respectively, were included.

Table 2 Adjusted clinical event rates for recurring events, by year after kidney transplantation.

Number of events per person-year	Year 1 of follow-up with leukopenia/neutropenia, n = 11524 (A)		Year 1 of follow-up without leukopenia/neutropenia, n = 3874 (B)		Relative risk A/B	Year 2-5 of follow-up with leukopenia/neutropenia, n = 55831 (A)		Year 2-5 of follow-up without leukopenia/neutropenia, n = 90351 (B)		Relative risk A/B
	Event rate (%)	SE (%)	Event rate (%)	SE (%)		Event rate (%)	SE (%)	Event rate (%)	SE (%)
CMV DNAemia1	3	0.037	1.8	0.083	1.67	1.25	0.061	0.69	0.084	1.81
CMV disease	0.09	0.001	0.02	0.001	3.95	0.06	0.001	0.01	0	5.14
Opportunistic infections (excluding CMV)	2.19	0.009	1.57	0.012	1.4	1.15	0.007	0.76	0.005	1.52
G-CSF use	0.9	0.004	0.24	0.003	3.76	0.27	0.003	0.09	0.001	3.04
All-cause rehospitalization	2.24	0.013	1.52	0.013	1.47	0.95	0.011	0.52	0.005	1.84

¹Only patients with available cytomegalovirus DNAemia lab data were considered in the denominator for the rate of this event. In year 1 follow-up, this included 652 patients with leukopenia/neutropenia and 106 patients without leukopenia/neutropenia. In years 2-5 of follow-up, this included 258 patients with leukopenia/neutropenia and 148 patients without leukopenia/neutropenia.

CMV: Cytomegalovirus; G-CSF: Granulocyte colony stimulating factor.

In year 1, patients with and without neutropenia and/or leukopenia had 3.00 vs 1.80 events, respectively, per person-year for CMV DNAemia (relative risk: 1.67) and 0.09 vs 0.02 events per person-year for CMV disease (relative risk: 3.95). Patients with neutropenia and/or leukopenia also had higher rates per person-year of opportunistic infections [2.19 vs 1.57 (relative risk 1.40)], G-CSF use [0.90 vs 0.24 (3.76)], and all-cause rehospitalization [2.24 vs 1.52 (1.47)] compared to patients without neutropenia and leukopenia. In years 2-5, the relative risk between the two cohorts showed a consistent trend where patients with neutropenia and/or leukopenia continued to experience a higher number of clinical events compared to patients without neutropenia and leukopenia.

One-time events: The rates of one-time clinical events are reported in Table 3. A total of 11997 (77.91%) patients with and 3401 (22.09%) without neutropenia and leukopenia were included in the analysis. In year 1, patients with neutropenia and/or leukopenia had higher event rates for acute graft rejection [21.14% vs 13.23% (relative risk: 1.60)], graft loss [11.56% vs 7.86% (1.47)], and all-cause mortality [0.06% vs 0.02% (2.55)] compared with patients without neutropenia and leukopenia. In years 2-5, patients with neutropenia and/or leukopenia continued to have higher event rates for acute graft rejection and graft loss compared to patients without neutropenia and leukopenia. The mortality event rates increased to approximately 9% for both cohorts and the relative risk was close to 1.

Table 3 Adjusted clinical event rates for one-time events, by year after kidney transplantation.

Proportion of patients with events	Year 1 of follow-up with leukopenia/neutropenia, n = 119971 (A)		Year 1 of follow-up without leukopenia/neutropenia, n = 34011 (B)		Relative risk A/B	Year 2-5 of follow-up with leukopenia/neutropenia2 (A)		Year 2-5 of follow-up without leukopenia/neutropenia2 (B)		Relative risk A/B
	Event rate (%)	SE (%)	Event rate (%)	SE (%)		Event rate (%)	SE (%)	Event rate (%)	SE (%)
Acute graft rejection	21.14	0.27	13.23	0.63	1.6	13.26	0.27	10.3	0.63	1.29
Graft loss	11.56	0.28	7.86	1.15	1.47	11.48	0.28	6.96	1.15	1.65
All-cause mortality	0.06	3.21	0.02	17.69	2.55	9.04	3.21	9.46	17.69	0.96

¹One-time events were evaluated over the entire duration of follow-up. Hence, the cohort of neutropenia or leukopenia was also defined using the lab data through the entire duration of follow-up. The n represented the number of patients at risk at day 0.

²The number of patients at risk after year 1 of follow-up among those with leukopenia/neutropenia by clinical event was: n = 9651 (acute graft rejection), n = 10685 (graft loss), and n = 11990 (all-cause mortality). The number of patients at risk after year 1 of follow-up among those without leukopenia/neutropenia by clinical event was: n = 2977 (acute graft rejection), n = 3141 (graft loss), and n = 3401 (all-cause mortality).

Patients with no events (excluding mortality): In year 1, 17.78% of patients with neutropenia and/or leukopenia were free from any non-fatal clinical events compared to 27.54% of patients without neutropenia and leukopenia (relative risk: 0.65); in years 2-5, the proportions were 18.88% vs 32.77% (relative risk: 0.58), respectively (Table 4).

Table 4 Adjusted proportion with no events except mortality, by year after kidney transplantation.

Proportion of patients with no event (except for mortality)	Year 1 of follow-up, n = 15398			Year 2-5 of follow-up, n = 14618
	With leukopenia/neutropenia, n = 11524 (A), %	Without leukopenia/neutropenia, n = 3874 (B), %	Relative risk A/B	With leukopenia/neutropenia, n = 5583 (A), %	Without leukopenia/neutropenia, n = 9035 (B), %	Relative risk A/B
No event	17.78	27.54	0.65	18.88	32.77	0.58

HCRU: HCRU excluding the index hospitalization is reported in Table 5. For the year 1 analyses of HCRU, 11516 (74.83%) patients with neutropenia and/or leukopenia and 3874 (25.17%) patients without neutropenia and/or leukopenia with a full year of continuous care were included; for years 2-5, 9264 (26.84%) and 25249 (73.16%) patient observations were included in the respective cohorts.

Table 5 Healthcare resource utilization by year after kidney transplantation.

All-cause healthcare resource utilization	Year 1, n = 153901		Year 2-5, n = 345132
	With leukopenia/neutropenia, n = 11516	Without leukopenia/neutropenia, n = 3874	With leukopenia/neutropenia, n = 9264	Without leukopenia/neutropenia, n = 25249
Inpatient
Patients with admission, n (%)	7441 (64.61)	2205 (56.92)	3408 (36.79)	6471 (25.63)
Number of admissions, mean ± SD (median)	3.47 ± 5.07 (2)	2.76 ± 3.63 (2)	2.70 ± 3.01 (2)	2.29 ± 2.57 (1)
Length of stay (days) among patients who had ≥ 1 admission, mean ± SD (median)	6.68 ± 27.26 (4)	7.55 ± 36.70 (4)	7.50 ± 29.83 (4.5)	5.92 ± 24.40 (3)
Emergency department/other
Patients with visits, n (%)	3383 (29.38)	988 (25.50)	1927 (20.80)	4036 (15.98)
Number of visits, mean ± SD (median)	1.81 ± 1.37 (1)	1.74 ± 1.43 (1)	1.87 ± 1.69 (1)	1.77 ± 1.65 (1)
Outpatient
Patients with visits, n (%)	11304 (98.16)	3813 (98.43)	9130 (98.55)	24460 (96.88)
Number of visits, mean ± SD (median)	48.97 ± 37.34 (37)	34.42 ± 28.18 (25)	31.73 ± 27.63 (25)	15.59 ± 18.53 (9)
Other3
Patients with other visits, n (%)	162 (1.41)	74 (1.91)	122 (1.32)	442 (1.75)
Number of other visits, mean ± SD (median)	2.85 ± 7.62 (1)	5.05 ± 12.84 (1.5)	3.25 ± 3.84 (2)	2.17 ± 2.11 (1)
Unknown
Patients with visits, n (%)	7702 (66.88)	2142 (55.29)	7473 (80.67)	14820 (58.70)
Number of visits, mean ± SD (median)	36.74 ± 31.57 (28)	28.27 ± 30.78 (16)	17.85 ± 18.30 (13)	13.48 ± 16.19 (8)
Any labs
Patients with any labs, n (%)	11475 (99.64)	3832 (98.92)	9264 (100.00)	25249 (100.00)
Number of any labs, mean ± SD (median)	1654.55 ± 984.16 (1485)	1182.27 ± 755.58 (1084)	622.37 ± 706.25 (419)	327.89 ± 374.36 (222)

¹The n represented the number of patients with a full year of follow-up during year 1.

²The n represented the number of patient observations with a full year of follow-up during years 2-5. Patients with multiple full years of follow-up were counted for each year.

³Other visits included home health visits, pre-admission encounters, and virtual visits.

Excluding the index hospitalization, 64.61% and 56.92% of the patients with and without neutropenia and leukopenia, respectively, had an inpatient admission in year 1 post-transplant, with an average of 3.47 and 2.76 admissions of 6.68 and 7.55 days, respectively among those with at least one inpatient admission. A total of 29.38% and 25.50% of the patients with and without neutropenia and/or leukopenia, respectively, had an ED visit, with an average of 1.81 and 1.74 visits among those who had at least one ED visit. Almost all patients in both cohorts (98.16% with vs 98.43% without neutropenia/leukopenia) had an outpatient visit in year 1, with an average of 48.97 and 34.42 visits, respectively, among those who had at least one outpatient visit. The average number of lab visits was 1643.55 and 1182.27 for patients with and without neutropenia and/or leukopenia, respectively, among patients with at least one lab visit which was almost all patients. In both cohorts in year 1, less than 2% of patients had other visits. Visits with unknown HCRU categories were identified in the data and more than half of the patients from both cohorts had at least one unknown HCRU visit.

The trend persisted in years 2-5 with patients experiencing neutropenia and/or leukopenia having a higher average number of HCRU visits compared to patients without neutropenia and leukopenia.

DISCUSSION

This retrospective cohort study described the short- and long-term clinical and HCRU burden among United States kidney transplant recipients with and without neutropenia and/or leukopenia who initiated valganciclovir and/or ganciclovir prophylaxis within 30 days after kidney transplantation. We found high risk of CMV disease and significant HCRU among this population. Within the first year post-kidney transplant among those initiated on valganciclovir and/or ganciclovir prophylaxis, approximately three-fourths of patients developed neutropenia and/or leukopenia.

Patients with neutropenia and/or leukopenia experienced higher rates of clinical events over both the short- and longer-term post-transplant periods than patients without neutropenia and leukopenia. During the first year following kidney transplantation, patients with neutropenia and/or leukopenia had 2-3 events each of CMV DNAemia, opportunistic infections, and all-cause hospitalization, and were 40%-70% more likely to have these clinical events than patients without neutropenia and leukopenia. Despite the absolute event rates for CMV disease and G-CSF use being lower than the other assessed clinical events, the likelihood of their occurrence in patients with neutropenia and/or leukopenia was 276%-295% higher than those without neutropenia and leukopenia. Acute graft rejection and graft loss were also 50%-60% more common in patients with than without neutropenia and leukopenia. The clinical burden also persisted up to five years in patients experiencing neutropenia and/or leukopenia after kidney transplantation and valganciclovir and/or ganciclovir treatment, underscoring the long-term elevation of risk.

Furthermore, patients with neutropenia and/or leukopenia had higher HCRU burden. On average, patients with neutropenia and/or leukopenia incurred 0.7 more inpatient admissions and 15 more outpatient visits during year 1 post-transplant than patients without neutropenia and leukopenia, a trend which extended up to five years after transplantation, with 0.4 more inpatient admissions and 16 more outpatient visits, among those who had at least one visit. Taken together, the higher rates of outpatient visits and inpatient admissions associated with neutropenia and/or leukopenia in this patient population translate to a greater burden on United States health systems, which could potentially be mitigated by improvements to CMV prophylaxis.

The findings of a recent systematic literature review (SLR) of observational studies on the burden of neutropenia and leukopenia among adult kidney transplant recipients are broadly consistent with those of the current study[21]. The SLR of 82 studies reported that, in the first year post-transplant, up to 48% of patients experienced neutropenia and up to 83% experienced leukopenia. The authors observed an increased risk of CMV infection and disease, acute rejection, graft failure, and mortality in patients with neutropenia and/or leukopenia, similar to the present findings. The economic burden of CMV disease has been established in existing literature. A United States claims study, which included 3,258 kidney transplant recipients during 2012-2018, reported that patients who developed neutropenia following valganciclovir prophylaxis incurred more HCRU, including ED visits, outpatient visits (dialysis and non-dialysis), inpatient hospitalizations, and medication use compared to patients without neutropenia[26]. The healthcare cost difference within the first year was estimated to be approximately 36000 USD per patient. The present findings build upon those in the prior literature to contribute new evidence regarding the substantial clinical and economic burden experienced by kidney transplant recipients with myelosuppressive conditions such as neutropenia or leukopenia after receiving standard of care anti-CMV therapy.

The 2020 Annual Data Report of kidney, developed by the United States Organ Procurement and Transplantation Network and the Scientific Registry of Transplant Recipients, identified a one-year death-censored graft failure rate of less than 5% for kidney transplant recipients, whether from living or deceased donors, and a one-year acute rejection rate of 6.8% in 2018 to 2019[27]. The current study has identified a higher one-year graft failure rate ranging from 8% to 12% and a higher one-year acute graft rejection rate ranging from 13% to 21%. While we intended to understand the discrepancies, factors that could impact transplantation success, such as recipient wait time, sensitization, mismatch (in terms of human leukocyte antigens and ABO blood group), quality of the donor kidney, warm and cold ischemic times, and the type of kidney (including considerations such as extended criteria, donation after cardiac death, donation after brain death, etc.) were not available in the current data. Furthermore, the study has identified an annual rate of 0.24 G-CSF use in patients without neutropenia and leukopenia during the first year following kidney transplantation. G-CSF might have been prescribed preventively in response to a potential decline in ANC in anticipation of neutropenia. Due to the nature of data extraction using natural language processing, there is a possibility of bias where G-CSF was noted but not actually consumed.

This study is subject to a few other limitations, some of which are inherent to retrospective studies using EMR data. First, the sample was derived from the TriNetX network, which may not be fully generalizable to the entire kidney transplantation population in the United States. Because TriNetX is not a closed system, prescriptions and encounters outside of the network might not be observed. This could result in missing data or underreporting of HCRU and clinical events. As the data cut intended to exclude patients who died within one year of kidney transplantation, this patient cohort may have lower disease severity and better observed outcomes than other populations. The mortality rates presented in this study should be interpreted with caution. Second, data abstraction relied on natural language processing, therefore the results could be impacted by any errors in data entry. Additionally, having a prescription record of valganciclovir or ganciclovir in the data may not indicate actual intake of the medication. Third, the available variables and level of detail were limited by what was reported in patients’ EMR and subsequently abstracted by TriNetX (i.e., detail for birth or death dates). Duration of valganciclovir and/or ganciclovir was not evaluated due to limited availability. Moreover, the study defined the neutropenia and/or leukopenia cohort during the same time frame as the outcome analysis, and patients could transition out of neutropenia or leukopenia during the outcome follow-up period. Therefore, the results for patients with neutropenia and/or leukopenia included time periods when patients might not have the conditions and the disease burden estimates were conservative. Lastly, since the treatment duration of valganciclovir and ganciclovir prophylaxis was not evaluated, caution is advised when interpreting the long-term outcomes in relation to the prophylactic treatments. Future research is recommended on alternative CMV prophylaxis methods with lower risks of neutropenia and/or leukopenia, as well as on their long-term impact the clinical outcomes of kidney transplant recipients.

CONCLUSION

Adults in the United States receiving valganciclovir and/or ganciclovir post-kidney transplantation who developed neutropenia and/or leukopenia experienced substantial clinical and economic burden, including more clinical events and higher HCRU up to five years post-transplantation, compared with patients who did not develop neutropenia and leukopenia. These findings may inform treatment decisions for clinicians and underscore the need for novel treatment options for CMV prophylaxis which are associated with lower risk of neutropenia and leukopenia to improve the outcomes of this patient population.

ACKNOWLEDGEMENTS

This study was supported by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, United States. The study sponsor was involved in several aspects of the research, including the study design, interpretation of data, and writing of the manuscript. Medical writing assistance was provided by Shelley Batts, PhD, an independent contractor of Analysis Group, Inc. and funded by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, United States.