Invasive fungal infections linked to sharply higher risks of death and graft failure after kidney transplantation

Invasive fungal infections after kidney transplantation are associated with dramatically higher risks of death or graft failure, and the danger is greatest in the first months after both transplantation and infection, according to a retrospective cohort study of 752 adult kidney transplant recipients published in Frontiers in Cellular and Infection Microbiology. The research, conducted at a single Chinese transplant center, found that recipients who developed an invasive fungal infection faced an incidence rate of the composite outcome of all-cause mortality or graft failure more than 34 times higher during infection-exposed periods within the first six months after transplantation than during unexposed periods, and nearly nine times higher beyond six months.

The study team, led by researchers at the Second Affiliated Hospital of Guangxi Medical University in Nanning, followed all adults who received kidney transplants at their center between 2020 and 2023, tracking outcomes through the end of December 2025. Of 1,525 transplantations performed during that window, 1,412 were in adults, and 752 recipients with reliable follow-up formed the final analysis cohort. Over a median follow-up of 937 days, 137 recipients, or 18.2 percent, developed a first invasive fungal infection, a rate of 75.2 infections per 1,000 person-years of observation.

The pathogen landscape in this cohort differed markedly from what has been reported in European and North American transplant populations. Pneumocystis jirovecii pneumonia, known as PJP, accounted for 99 of the 137 infections, or 72.3 percent, while invasive aspergillosis accounted for 29 cases, or 21.2 percent. Invasive candidiasis, cryptococcosis, and a single mucorales infection made up the remainder. Nearly all infections, 95.6 percent, involved the lungs. The median time from transplantation to first infection was 284 days, with 51 episodes occurring within six months of surgery and 86 occurring later.

Diagnoses were not taken from administrative codes, a limitation the authors identified in much of the prior literature. Instead, each candidate case was adjudicated by transplant physicians using the 2020 consensus definitions of the European Organisation for Research and Treatment of Cancer and the Mycoses Study Group Education and Research Consortium, integrating host factors, clinical and radiographic findings, and microbiological evidence. Because Pneumocystis PCR was not routinely performed at the center, supportive evidence for PJP included detection of the organism by metagenomic next-generation sequencing of respiratory specimens or at least two consecutive elevated serum beta-D-glucan measurements, always interpreted alongside the full clinical picture. Only proven or probable infections counted; colonization and isolated positive tests without a compatible syndrome were excluded.

The analytical design was built to preserve the correct ordering of cause and outcome. Infection status was treated as a time-varying exposure, so recipients contributed unexposed person-time before diagnosis and exposed person-time afterward, preventing the misclassification of periods before infection as exposed. A continuous-time Cox model with a spline interaction allowed the infection-associated hazard ratio to change over post-transplant time, and it showed a statistically significant decreasing pattern, with a P value for time variation of 0.002. During the first six months, the incidence rate of the composite outcome was 1.792 events per person-year during infection-exposed time versus 0.053 during unexposed time, an incidence rate ratio of 34.05. Beyond six months, the corresponding rates were 0.138 and 0.016, an incidence rate ratio of 8.80. Both comparisons were highly statistically significant.

To validate these findings temporally, the researchers used sequential risk-set matching. Each infection case was matched to four control records from recipients transplanted within 30 days of the case, alive with a functioning graft at the case’s infection date, and not yet infected. This produced 137 matched sets containing 137 case records and 548 control records, all assigned the same index date. Within one year of that index date, the composite outcome occurred in 37 infected records and 17 control records, with cumulative probabilities of 27.5 percent and 3.3 percent respectively. The overall hazard ratio for the composite outcome was 11.20, but the association was heavily front-loaded: 21.19 during the first 90 days after the index date, falling to 2.72 between days 91 and 365.

Secondary outcomes told a consistent story. In the matched cohort, infection was associated with all-cause mortality over one year with a hazard ratio of 13.15, again strongest in the first 90 days at 21.42. Death-censored graft failure showed an overall hazard ratio of 6.83, with a point estimate of 20.00 in the first 90 days, though that early estimate was imprecise given sparse events. Acute rejection, an outcome rarely examined after fungal infection in prior work, was also more frequent after infection, with an overall hazard ratio of 3.89. Complementary Fine-Gray subdistribution hazard models, which treat mortality and graft failure as competing events where appropriate, produced broadly similar estimates, including a subdistribution hazard ratio of 13.24 for mortality and 3.40 for rejection. In a multivariable model, the infection association persisted with an adjusted hazard ratio of 10.83, while a history of diabetes and a respiratory bacterial infection in the preceding 180 days were also independently associated with the composite outcome.

Within the infected group itself, timing and pathogen type mattered. Recipients whose infection began within six months of transplantation had a one-year cumulative probability of the composite outcome after diagnosis of 41.3 percent, compared with 19.0 percent for those infected later, an unadjusted hazard ratio of 2.38 for early versus late onset. All-cause mortality followed the same pattern with a hazard ratio of 2.34. Comparing pathogens head-to-head, the contrast was starker still: recipients with invasive aspergillosis had a one-year cumulative probability of the composite outcome of 55.9 percent, versus 17.4 percent for those with PJP, a hazard ratio of 0.25 when comparing PJP with aspergillosis. Mortality showed the same direction, and laboratory measurements around the time of diagnosis showed that recipients with aspergillosis had lower observed estimated glomerular filtration rates, lymphocyte counts, hemoglobin, and platelet counts than those with PJP.

The authors caution that these are observational associations, not causal effects. Residual confounding cannot be excluded, and the exploratory within-infection comparisons were unadjusted, so the apparent advantage of PJP over aspergillosis could reflect differences in infection severity, timing, or underlying recipient fragility rather than the pathogen itself. The single-center retrospective design also carries specific limitations: 660 recipients were excluded because they discontinued follow-up within the first year without a documented death or graft failure, a group whose outcomes could not be ascertained and whose exclusion may have shaped the observed infection rate. Fungal testing was performed only when clinically indicated rather than through routine screening, raising the possibility that subclinical infections went undetected. Individual adherence to prophylaxis with trimethoprim-sulfamethoxazole and ganciclovir could not be reliably reconstructed from medical records and was not analyzed.

The unusually high infection rate and PJP dominance in this cohort also deserve context. A recent prospective five-center German cohort of 1,258 kidney recipients reported a 6.7 percent cumulative incidence of fungal infection, with Candida and Aspergillus predominating and PJP accounting for only 13.5 percent of cases. The authors suggest that differences in diagnostic practices, including their use of metagenomic sequencing as supportive evidence for PJP, may partly explain the gap, alongside the exclusion of recipients lost to follow-up. They emphasize that the findings warrant multicenter validation in cohorts with different pathogen distributions and diagnostic approaches. In the meantime, the results support coordinated attention to infection control, graft function, and immunological monitoring after a fungal infection diagnosis, with the timing of onset and the infecting organism both shaping the risk landscape that recipients and their clinicians face.

Subject of Research: Longitudinal association of invasive fungal infections with mortality, graft failure, and rejection after kidney transplantation

Article Title: Do invasive fungal infections affect kidney transplantation outcomes?

Article References: Ma, R., Zhang, D., Luo, Z., Sun, J., Sun, N., Liang, C., Hu, C., Li, H., Lan, L., Wu, J., Dong, J., Wen, J., Sun, X., & Wen, N. (2026). Do invasive fungal infections affect kidney transplantation outcomes?. Frontiers in Cellular and Infection Microbiology, 16, 1982586. https://doi.org/10.3389/fcimb.2026.1982586

Image Credits: AI Generated

DOI: 10.3389/fcimb.2026.1982586

Keywords: kidney transplantation, invasive fungal infection, Pneumocystis jirovecii pneumonia, invasive aspergillosis, graft failure, all-cause mortality, acute rejection, time-varying analysis, risk-set matching, retrospective cohort study, immunosuppression, transplant outcomes