Monthly variance in UK renal transplantation activity: a national retrospective cohort study

Objective To identify whether renal transplant activity varies in a reproducible manner across the year. Design Retrospective cohort study using NHS Blood and Transplant data. Setting All renal transplant centres in the UK. Participants A total of 24 270 patients who underwent renal transplantation between 2005 and 2014. Primary outcome Monthly transplant activity was analysed to see if transplant activity showed variation during the year. Secondary outcome The number of organs rejected due to healthcare capacity was analysed to see if this affected transplantation rates. Results Analysis of national transplant data revealed a reproducible yearly variance in transplant activity. This activity increased in late autumn and early winter (p=0.05) and could be attributed to increased rates of living (October and November) and deceased organ donation (November and December). An increase in deceased donation was attributed to a rise in donors following cerebrovascular accidents and hypoxic brain injury. Other causes of death (infections and road traffic accidents) were more seasonal in nature peaking in the winter or summer, respectively. Only 1.4% of transplants to intended recipients were redirected due to a lack of healthcare capacity, suggesting that capacity pressures in the National Health Service did not significantly affect transplant activity. Conclusion UK renal transplant activity peaks in late autumn/winter in contrast to other countries. Currently, healthcare capacity, though under strain, does not affect transplant activity; however, this may change if transplantation activity increases in line with national strategies as the spike in transplant activity coincides with peak activity in the national healthcare system.


GENERAL COMMENTS
The authors intend to address two different questions: 1.) Does organ donation and renal transplantation show seasonal variance in the UK 2.) Are (kidney) transplant rates affected by healthcare capacity Ad 1.) As far as can be judged by the Methods and Results section, donor rates for the different causes of death were based on the number of utilized donors in the different months. It would be important to look at the number of reported donors for each of these categories, because differences in overall acceptance rates in the different months could influence this finding. Possible confounding factors like donor age distribution over the different months (reported donors, utilized donors) should be looked at to see, whether this affects the conclusions.

Ad 2.)
The analysis of the influence of healthcare capacity is dependent on the reliability of the data provided by the transplant centers regarding the number of turndowns based on lack of capacity. If a center reports a high rate of turndowns for this reason, patients, referring doctors and authorities might lose trust in this center. Therefore, it seems likely that underreporting of this reason for turning down an organ offer took place. It could be possible, that the kidneys of a donor were officially not accepted for "donor quality" reasons but in fact, "transplant capacity" was not available. Are there any data on the number of organs that were turned down due to poor organ quality by one center and were then transplanted in another center?
These points should addressed to better understand, whether the conclusions of this retrospective study are justified.
Additional specific points: Page 6/23, line 18: The authors claim that there was an increase in patient mortality and morbidity in recent years. A reference for this statement should be provided and it should be clarified whether this statement refers to patients on the waiting list or patients with ESRD.
Page 8/23, line 30: It should read, "Deceased donor transplants form about two thirds of all UK kidney transplant activity…" The figures are small and difficult to read, comparison between transplant and donation rates are difficult to make. The authors should consider adding a figure that shows the relative contribution of the different causes of donor death to the overall number of donors in the different months.

GENERAL COMMENTS
@Overall # This is an interesting study investigating the seasonal variations in renal transplant activities in the UK between 2004 and 2014. # The study is well-conducted and written and the results are of some importance to policy makers. However, there are some minor corrections that may make the study better, in my opinion.
@ Results and data analysis # The statistical method used in this study is adequate and convincing; however, the authors did not report the results in detail; i.e. the GEEs and GLM models. The readers may need additional information to better understand the results. Perhaps, the authors did not reported the results in detail due to word count limits, so I suggest adding more details about the results as supplementary materials.
@ Limitations # The study uses aggregate data for all renal transplant activities in the UK; however, the results of this study (seasonal variations) may not necessarily be true for some transplant centers. Therefore, at the micro-level (transplant center), the results of this study would be useful, but further center-specific analysis is also important. The authors need to address this issue as a limitation of this study.
@ Writing and reporting # The article is well written, however, there are some missed commas in the introduction, in the following sentences: "Despite recent improvements in the United Kingdom (UK) and other countries regarding access to transplantation demand for renal transplantation exceeds the number of available donors leading to increased patient mortality and morbidity." "Since many potential donors cannot be utilised for various reasons it is vital that these are minimised so that every potential organ for transplant is utilised for the primary intended recipient." # The comma in the following sentence should be deleted: "The other one third of kidney transplants, are transplant from living donors." We agree that the donor rates reported in the paper are from actual transplants that have been carried out and therefore total donor offer rates were not examined. We also agree that changes in donor acceptance rates could be a potential explanation for our findings if we only studied one year. Our cohort however was over ten years with the pattern staying the same for each individual year during this period making it very unlikely that donor acceptance rates skewed our findings. Furthermore, other epidemiological studies have confirmed the winter spike in incidence for the main cause of donors i.e. hypoxic brain injury or intracranial events as mentioned in the introduction Ad 2.) The analysis of the influence of healthcare capacity is dependent on the reliability of the data provided by the transplant centers regarding the number of turndowns based on lack of capacity. If a center reports a high rate of turndowns for this reason, patients, referring doctors and authorities might lose trust in this center. Therefore, it seems likely that underreporting of this reason for turning down an organ offer took place. It could be possible, that the kidneys of a donor were officially not accepted for "donor quality" reasons but in fact, "transplant capacity" was not available. Are there any data on the number of organs that were turned down due to poor organ quality by one center and were then transplanted in another center?

REVIEWER
We do agree that a potential limitation of any study, not only ours, is the reliability of data. To ensure that our data was robust as possible it was collected by a nationally recognised body, independent of the investigators, furthermore where a transplant centre declines a donor to a named recipient the data is audited by NHSBT. We now make this clearer vy changing the limitations of the study to "The national transplant database is filled using data submitted by each transplant center; therefore the data has not been independently corroborated." Even if we had the data on organs that have been turned down by one centre but used in another centre it would be difficult to separate out the nuances associated with the organ declines clearly establishing that the reason was due to lack of capacity. This is because each centre accepts different quality of organs depending on local expertise therefore an organ that is deemed unsuitable at one centre may be suitable at another.
Ultimately transplant teams desire what is best for the donor and recipient i.e. a successful transplant and robust reporting of outcomes is an important part of this process.
These points should addressed to better understand, whether the conclusions of this retrospective study are justified.
Additional specific points: Page 6/23, line 18: The authors claim that there was an increase in patient mortality and morbidity in recent years. A reference for this statement should be provided and it should be clarified whether this statement refers to patients on the waiting list or patients with ESRD.
This statement was intended to say that since demand is higher than supply, not everyone will get a kidney transplant who needs one. This therefore results in increased patient morbidity and mortality. This has been updated with the correct references in the manuscript.
Page 8/23, line 30: It should read, "Deceased donor transplants form about two thirds of all UK kidney transplant activity…" We agree and this has now been corrected The figures are small and difficult to read, comparison between transplant and donation rates are difficult to make. The authors should consider adding a figure that shows the relative contribution of the different causes of donor death to the overall number of donors in the different months.
We did consider doing this however due to concerns about presenting the same data twice, we feel it is best to present the data in its current format since this allows the reader to clearly see temporal changes in activity per cause of donor death. However the requested graph is presented below for the reviewers information: Please leave your comments for the authors below @Overall # This is an interesting study investigating the seasonal variations in renal transplant activities in the UK between 2004 and 2014. # The study is well-conducted and written and the results are of some importance to policy makers. However, there are some minor corrections that may make the study better, in my opinion.
@ Results and data analysis # The statistical method used in this study is adequate and convincing; however, the authors did not report the results in detail; i.e. the GEEs and GLM models. The readers may need additional information to better understand the results. Perhaps, the authors did not reported the results in detail due to word count limits, so I suggest adding more details about the results as supplementary materials.
In order to achieve this we have now uploaded the statistical code into a data repository. If readers would like to know more information, then of course they can contact either corresponding author who will provide the information.
@ Limitations # The study uses aggregate data for all renal transplant activities in the UK; however, the results of this study (seasonal variations) may not necessarily be true for some transplant centers. Therefore, at the micro-level (transplant center), the results of this study would be useful, but further center-specific analysis is also important. The authors need to address this issue as a limitation of this study.
This has now been added to the manuscript as below: It was therefore reassuring that we found at the national level no solid evidence of transplant surgery utilising deceased donors being cancelled during the winter months. Despite this however it is important that transplant departments plan for this predictable and reproducible surge in transplant activity making sure their own activity is not affected.
@ Writing and reporting # The article is well written, however, there are some missed commas in the introduction, in the following sentences: "Despite recent improvements in the United Kingdom (UK) and other countries regarding access to transplantation demand for renal transplantation exceeds the number of available donors leading to increased patient mortality and morbidity." "Since many potential donors cannot be utilised for various reasons it is vital that these are minimised so that every potential organ for transplant is utilised for the primary intended recipient." # The comma in the following sentence should be deleted: "The other one third of kidney transplants, are transplant from living donors." We agree and thank the reviewer for pointing this out. It has now been corrected in the manuscript. I would like the authors to address the issues below: 1) LD rate or variationdue to intrinsic donor/patient factors or factors within the healthcare system? Surgery availability (school holidays and hence surgical availability). Do the authors have donor relationship to recipients?
Unfortunately, we do not have access to this data and this is one of the reasons why we state in the discussion that a detailed causal analysis is beyond the scope of this manuscript. Since the paper is focused on establishing whether any significant variation exists rather than establishing a causal relationship we do not feel the lack of this analysis subtracts from the message of the overall paper.
2) Did (or can) the authors account for centre effect or era variations (patterns may have changed over time especially changes in donor characteristics and acceptance criteria) We agree that these are important points and have tried to address them below: There are centre variations in the acceptance of higher donor risk index kidneys and the ration of DBD and DCD kidneys accepted. These are demonstrated very clearly in the annual centre specific reports produced by NHSBT. Despite this centre variation it cannot explain the national variation reported in the manuscript since all rejected kidneys are offered nationally, allowing other centres to accept them for transplantation.
Era variations were accounted for by creating ten separate GLMs, one for each year under investigation. No significant variations in patterns between these models were found and we have made this clearer in the methods section: "To account for the unknown correlation structure in our data, generalised estimating equations (GEEs) per year were used to estimate the GLM parameters" 3) Can the authors comment further on the variation in infection deaths (DD)cause of death recorded?
Infectious deaths were due to meningitis (n=280), pneumonia (n=61), septicaemia (n=17) and were unspecified in nine cases. The small numbers preclude any robust seasonal variation for the subtypes.

4) What about DCD donors/transplant activity?
We thank the reviewer for their suggestion and have performed the suggested analysis. This shows that seasonal variation mainly affects donation after brain death rather than donation after circulatory arrest. The results are presented in two additional tables (table 2 and 3) and the following sentence has been added to the text in the section titled seasonal variance.
"This was mainly seen in donation after brain death (table 2) rather than circulatory death (table 3)."