Unveiling Donor-Derived BKPyV DNAemia Through Analysis of Contralateral Kidney Transplant Recipients

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Review Biomedicines

 

This paper by Moest et al investigates the impact of donor-derived BKPyV infections in kidney transplant patients. BKPyV infections is a notable concern in kidney transplant patients since a significant number of the recipients are unfortunately expected to achieve BKPyV-associated nephropathy. The investigators approach is to retrospectively characterise whether there is any correlation between recipients who are given kidneys from the same donor. Their result implies that the frequency of BKPyV infections is almost three times higher where the recipients of the donor are BKPyV-positive compared to recipients of the donors are BKPyV-negative. The topic is of great concern and their design and the scientific design/approach is highly relevant although some points raise concern (see points 1-4 below). In addition, they characterised and compared some clinical parameters and although this is also relevant and potentially interesting, the relatively small sizes of the cohorts reduce the power of these comparisons.

 

Major points

1. My main concern with current study is their study design, where 117 BKPyV- positive patients from 2011-2021 were matched with 133 BKPyV-negative patients from 2018-2021. To increase the validity and reduce the risk for unwanted biases, I would prefer unselected cohorts from a defined period, i.e. all patients from a given period should have been included, irrespectively whether they were BKPyV-positive or not. With the design used in the paper, it cannot be excluded that the treatment and outcome varied during the period, which would have different consequences of the cohorts.

 

2. What is the expected frequency of patients achieving BKPyV DNAemia during the first year in this hospital setting and has it changed during the study period? In the context of the this would be a relevant piece of information.

 

3. In the presented data, the frequency of BKPyV DNAemia is significantly lower in the contralateral recipients compared to the selected recipients (22/117 versus 52/117). Are there any conceivable explanations for this difference and are there any risk for downstream consequences on the study?

 

4. In the study, chi-square analysis is consistently used for analysing statistical significance of categorical data. Are there any reasons why this method is used rather than Fisher´s exact test, which in contrast to chi-square is exact and has higher stringency? Although chi-square was commonly used in the precomupter era due to simpler maths, there are now web-based softwares for Fisher´s exact test for freely available.

 

 

5. In Table 1, the analysis of statistical significance is done for groups of parameters rather than each one individually (“Induction” and “Maintenance immunosuppression”). Are there any particular reason for performing this in groups rather than individually? By looking at the data, it appears that only a few of the individual parameters could have contributed to the statistical significance?

 

Minor points

1. According to line 226-227, BKPyV DNAemia was more frequently initiated on everolimus-based immunosuppression in the contralateral group. Although this is formally is correct, this small difference does not appear to be significant whereas other data in the table seems to be of higher significance and more relevant to mention in the text.

 

 

 

 

 

Author Response

Reviewer 1

Q1.My main concern with current study is their study design, where 117 BKPyV- positive patients from 2011-2021 were matched with 133 BKPyV-negative patients from 2018-2021. To increase the validity and reduce the risk for unwanted biases, I would prefer unselected cohorts from a defined period, i.e. all patients from a given period should have been included, irrespectively whether they were BKPyV-positive or not. With the design used in the paper, it cannot be excluded that the treatment and outcome varied during the period, which would have different consequences of the cohorts.

We thank the reviewer for this important comment. We acknowledge that including all transplant recipients from a single, predefined period would be the most unbiased approach. However, given the relatively low incidence of BKPyV DNAemia, such a design would have required inclusion of a substantially larger cohort, which was not feasible within the scope of this study.

To address the potential for time-related bias, we performed a sensitivity analysis restricted to recipients transplanted between 2018 and 2021, representing all patients with a postmortal kidney transplantation within the LUMC in this time period. This analysis yielded consistent results (See the section ‘sensitivity analysis’), supporting our findings. This is no more clearly noted in the methods section.

We agree that changes in clinical practice over time may have influenced treatment and outcomes, and this limitation has been acknowledged in the discussion. 

Q2.What is the expected frequency of patients achieving BKPyV DNAemia during the first year in this hospital setting and has it changed during the study period? In the context of the this would be a relevant piece of information.

In a previously described cohort from our center, the incidence of BKPyV DNAemia within the first year after transplantation remained stable over time. Specifically, the incidence was 19.9% (74/372) in 2011–2013, 20.0% (89/446) in 2014–2017, and 22.9% (59/258) in 2018–2020 (p=0.595), https://doi.org/10.1002/jmv.70084. Supporting the assumption that calendar-time effects did not meaningfully influence the observed associations. This has now been added to the discussion.

“ In addition, in a previously described cohort from our center, the incidence of BKPyV DNAemia within the first year after transplantation remained stable over time (2011–2013: 19.9%, 2014–2017: 20.0%, 2018–2020: 22.9%; p=0.595) supporting the assumption that calendar-time effects did not meaningfully influence the observed associations.”

Q3. In the presented data, the frequency of BKPyV DNAemia is significantly lower in the contralateral recipients compared to the selected recipients (22/117 versus 52/117). Are there any conceivable explanations for this difference and are there any risk for downstream consequences on the study?

The apparent difference in BKPyV DNAemia between contralateral and selected recipients reflects the patient selection and follow-up process, not a biological effect. We have now tried to further clarify this distribution/method in Figure 1, which shows the final cohort and the occurrence of BKPyV DNAemia among both index and contralateral recipients

 Q4.In the study, chi-square analysis is consistently used for analysing statistical significance of categorical data. Are there any reasons why this method is used rather than Fisher´s exact test, which in contrast to chi-square is exact and has higher stringency? Although chi-square was commonly used in the precomupter era due to simpler maths, there are now web-based softwares for Fisher´s exact test for freely available.

We thank the reviewer for this comment. We have now applied Fisher’s exact test for variables with small cell counts (<5: e.g., PRA, induction therapy, maintenance immunosuppression, and BK nephropathy). This did not materially change the results or their statistical significance. Table 1. and Methods section has been updated accordingly.

 Q5. In Table 1, the analysis of statistical significance is done for groups of parameters rather than each one individually (“Induction” and “Maintenance immunosuppression”). Are there any particular reason for performing this in groups rather than individually? By looking at the data, it appears that only a few of the individual parameters could have contributed to the statistical significance?

In Table 1, immunosuppressive regimens were analyzed in grouped categories to ensure sufficient numbers per group for meaningful statistical testing. While it is true that only a few individual regimens may drive the overall differences, separate testing of each regimen would result in very small cell counts and potentially unreliable statistics.

Importantly, the associations between immunosuppressive regimens and BKPyV DNAemia were further assessed and corrected in the Cox regression analyses, which consider individual risk factors.

 Q6. According to line 226-227, BKPyV DNAemia was more frequently initiated on everolimus-based immunosuppression in the contralateral group. Although this  formally is correct, this small difference does not appear to be significant whereas other data in the table seems to be of higher significance and more relevant to mention in the text.

We agree with the reviewer. To avoid overemphasis on a minor observation, we have revised the text to focus on the more relevant and statistically significant differences presented in Table 1.

Reviewer 2 Report

Comments and Suggestions for Authors

I read with great interest the article of Moest et al. on “Unveiling Donor-Derived BKPyV DNAemia through Analysis of Contralateral Kidney Transplant Recipients”.

The study focuses on donor-derived BKPyV in recipients of deceased donors. They conducted a retrospective, multi-center cohort study, evaluating the risk of developing BKPyV DNAemia in kidney recipient pairs from the same deceased donor.

The objective is clear and the study was conducted properly, even if the originality and/or size of the cohort do not provide new data.

However, several questions arise:

- Given the (validated) hypothesis that the virus originates from the recipient, why exclude patients who develop BKPyV DNAemia after treatment for rejection (line 82)?

- (line 159-164): Patients are selected. On one hand, those with BKPyV DNAemia during the 2010–2021 period, and on the other, those without DNAemia between 2018 and 2021. Why not use the same time periods to strengthen the data and avoid bias? Indeed, haven’t there been changes in BK virus screening strategies during this long period? Hasn’t the epidemiology changed?

- -Is urinary testing not available because some transplant recipients begin to replicate BKPyV, which is detectable in urine, and mount a rapid immune response without the viral genome being detected in the blood? This phenomenon is therefore highly dependent on the recipient and not on the donor. Only urine samples would allow for a truly reliable interpretation.

-  Do you have any sequence data between the positive pairs that could help us take the analysis further ?

Author Response

Reviewer 2

Q1.  Given the (validated) hypothesis that the virus originates from the recipient, why exclude patients who develop BKPyV DNAemia after treatment for rejection (line 82)?

We excluded patients who developed BKPyV DNAemia following rejection therapy to minimize confounding by intensification of immunosuppression. As rejection episodes are commonly treated with high-dose corticosteroids and/or IL-2 receptor–targeting therapies, these therapies are known to increase the risk of BKPyV reactivation (https://doi.org/10.1097/tp.0000000000004976 ). The sentence in the article is now extended to:” Patients who developed BKPyV DNAemia following rejection therapy were excluded to minimize confounding by increased immunosuppression”.

Q2. (line 159-164): Patients are selected. On one hand, those with BKPyV DNAemia during the 2010–2021 period, and on the other, those without DNAemia between 2018 and 2021. Why not use the same time periods to strengthen the data and avoid bias? Indeed, haven’t there been changes in BK virus screening strategies during this long period? Hasn’t the epidemiology changed?

We acknowledge that including all transplant recipients from a single, predefined period would be the most unbiased approach. However, given the relatively low incidence of BKPyV DNAemia, such a design would have required a substantially larger cohort, which was not feasible within the scope of this study. To address the potential for time-related bias, we performed a sensitivity analysis restricted to recipients transplanted between 2018 and 2021, representing all postmortem kidney transplants at the LUMC during this period. This analysis yielded consistent results (see ‘Sensitivity analysis’ section), supporting our findings.

In addition, in a previously described cohort from our center, the incidence of BKPyV DNAemia within the first year after transplantation remained stable over time (2011–2013: 19.9%, 2014–2017: 20.0%, 2018–2020: 22.9%; p=0.595), supporting the assumption that calendar-time effects did not meaningfully influence the observed associations.

This has been addressed in the discussion as well and the following text was added to the discussion:

“ In addition, in a previously described cohort from our center, the incidence of BKPyV DNAemia within the first year after transplantation remained stable over time (2011–2013: 19.9%, 2014–2017: 20.0%, 2018–2020: 22.9%; p=0.595) supporting the assumption that calendar-time effects did not meaningfully influence the observed associations.”

 

Q3. Is urinary testing not available because some transplant recipients begin to replicate BKPyV, which is detectable in urine, and mount a rapid immune response without the viral genome being detected in the blood? This phenomenon is therefore highly dependent on the recipient and not on the donor. Only urine samples would allow for a truly reliable interpretation.

Unfortunately, urinary BKPyV data were not systematically available, as urine testing is not part of routine screening in the participating centers. This has now been added to the limitation section.

Q4.  Do you have any sequence data between the positive pairs that could help us take the analysis further ?

Unfortunately, no suitable samples were available for retrospective analysis. Such analyses could have provided additional insight into donor-derived transmission and would be of interest for future studies. This limitation has now been explicitly acknowledged in the revised manuscript.

 

Reviewer 3 Report

Comments and Suggestions for Authors

This study on BKPyV further supports a likely donor–recipient link in BK polyomavirus reactivation after kidney transplantation. The authors show that when one recipient develops BKPyV DNAemia within the first-year post-transplant, the recipient of the contralateral kidney is at significantly higher risk of also developing BKPyV DNAemia (HR = 4.9). This is an interesting retrospective analysis conducted in recipients of kidneys from deceased donors and provides clinically relevant insight into the possible donor origin of post-transplant BKPyV replication.

 

To further strengthen the hypothesis that the replicating virus observed in both recipients is indeed the same viral strain, it would have been highly informative to perform viral genotyping and to analyze possible rearrangements in the NCCR region. Such analyses could provide additional molecular support for donor-derived transmission or shared viral origin. At the very least, the absence of these data should be acknowledged more explicitly in the limitations section.

In the Discussion, the authors mention vaccination strategies. It would also be valuable to discuss the potential role of preventive monoclonal neutralizing antibodies directed against all BKPyV serotypes, especially given that at least one such antibody is currently under development.

In addition, the authors may wish to cite and discuss the interesting work from the Strasbourg group (Solis et al.), which explored neutralizing antibody thresholds against the corresponding BKPyV serotype as predictors of protection against viral replication. This is particularly relevant to the present study and could enrich the discussion on donor/recipient immune status and risk stratification. If serum samples are available from the transplant recipients included in this cohort, a seroneutralization assay could also be considered as a valuable complementary analysis.

Finally, are biopsy samples available from these patients, either at the time of transplantation or during post-transplant follow-up? Assessing the presence of BKPyV in allograft biopsies could provide additional insight into the timing and tissue-level evidence of viral transmission and/or replication.

Author Response

Q1. To further strengthen the hypothesis that the replicating virus observed in both recipients is indeed the same viral strain, it would have been highly informative to perform viral genotyping and to analyze possible rearrangements in the NCCR region. Such analyses could provide additional molecular support for donor-derived transmission or shared viral origin. At the very least, the absence of these data should be acknowledged more explicitly in the limitations section.

Unfortunately, no suitable samples were available for retrospective analysis. Such analyses could have provided additional insight into donor-derived transmission and would be of interest for future studies. This limitation has now been explicitly acknowledged in the revised manuscript.

“In addition, no suitable samples were available for viral sequencing, preventing further molecular confirmation of whether the viruses in recipient pairs were identical.”

Q2. In the Discussion, the authors mention vaccination strategies. It would also be valuable to discuss the potential role of preventive monoclonal neutralizing antibodies directed against all BKPyV serotypes, especially given that at least one such antibody is currently under development.

Suggestion of the reviewer has been added to the discussion.

“Alongside vaccination, broadly neutralizing and monoclonal antibodies targeting BKPyV are currently being explored as both therapeutic and pre-emptive strategies. Whether these approaches are effective will need to be determined in ongoing and future studies.”

Q3. In addition, the authors may wish to cite and discuss the interesting work from the Strasbourg group (Solis et al.), which explored neutralizing antibody thresholds against the corresponding BKPyV serotype as predictors of protection against viral replication. This is particularly relevant to the present study and could enrich the discussion on donor/recipient immune status and risk stratification. If serum samples are available from the transplant recipients included in this cohort, a seroneutralization assay could also be considered as a valuable complementary analysis.

As mentioned, no suitable samples were available for retrospective analysis. We agree that such an additional analysis would have been valuable. The work of solis et al has been added to the discussion.

“as Solis et al. have shown that mismatches between genotype-specific neutralizing antibodies and the corresponding BKPyV strain are associated with an increased risk of BKPyV-related infection, whereas the presence of high neutralizing antibody titers against the specific BKPyV genotype confers protection”

Q4. Finally, are biopsy samples available from these patients, either at the time of transplantation or during post-transplant follow-up? Assessing the presence of BKPyV in allograft biopsies could provide additional insight into the timing and tissue-level evidence of viral transmission and/or replication.

Biopsy samples were not systematically collected, as protocol biopsies were not routinely performed. Consequently, assessment of BKPyV in allograft tissue could not be undertaken in this study.

 

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

All my concerns have been properly addressed and I have no further comments.