Pediatric deceased donor kidney transplant outcomes under the Kidney Allocation System

Abstract

The Kidney Allocation System (KAS) has resulted in fewer pediatric kidneys being allocated to pediatric deceased donor kidney transplant (pDDKT) recipients. This had prompted concerns that post-pDDKT outcomes may worsen. To study this, we used SRTR data to compare outcomes of 953 pre-KAS pDDKT (age<18 years) recipients (12/4/2012–12/3/2014) to 934 post-KAS pDDKT recipients (12/4/2014–12/3/2016). We analyzed mortality and graft loss using Cox regression, delayed graft function (DGF) using logistic regression, and length of stay (LoS) using negative binomial regression. Post-KAS recipients had longer pre-transplant dialysis times (median 1.26 vs. 1.07 years, p=0.02) and were more often cPRA 100% (2.0% vs. 0.1%, p=0.001). Post-KAS recipients had less graft loss than pre-KAS recipients (hazard ratio [HR]:_0.350.54_0.83, p=0.005), but no statistically significant differences in mortality (HR:_0.290.72_1.83, p=0.5), DGF (odds ratio [OR]:_0.931.32_1.93, p=0.2), and LoS (LoS ratio:_0.961.06_1.19, p=0.4). After adjusting for donor/recipient characteristics, there were no statistically significant post-KAS differences in mortality (adjusted HR [aHR]:_0.371.04_2.92, p=0.9 ), DGF (adjusted OR:_0.941.41_2.13, p=0.1), or LoS (adjusted LoS ratio:_0.931.04_1.16, p=0.5). However, post-KAS pDDKT recipients still had less graft loss (aHR:_0.380.59_0.91, p=0.02). KAS has had a mixed effect on short-term post-transplant outcomes for pDDKT recipients, although our results are limited by only two years of post-transplant follow-up.

INTRODUCTION

The Kidney Allocation System (KAS) was implemented on December 4, 2014, representing the largest change to deceased donor kidney allocation policy in the United States in over twenty years. The goals of KAS were to improve access to deceased donor kidney transplantation (DDKT) for historically disadvantaged groups (such as racial minorities and highly sensitized patients) and to better match the highest quality donor organs with recipients who have the longest expected post-transplant survival (1). Except for regional and national sharing priority for highly sensitized candidates, pediatric recipients’ place in the allocation sequence was not directly modified by KAS. However, a significant change was made to the method in which kidneys were allocated to them (2). Prior to KAS, pediatric deceased donor kidney transplant (pDDKT) recipients were allocated kidneys under Share-35, whereby kidneys from donors less than 35 years old would be initially allocated to pDDKT (3). Under KAS, pediatric candidates are now preferentially allocated kidneys with a Kidney Donor Profile Index (KDPI) < 35, representing a donor organ that is similar to the top 35% of kidneys recovered in the previous year (4). The goal of these changes was to maintain the pediatric priority for higher quality donor organs. Although early simulations suggested that pDDKT would remain unchanged after KAS implementation, changes in the allocation system can produce unintended consequences (2, 5).

Early reports on the effect of KAS were encouraging, with studies showing an increase in transplantation rates for highly sensitized patients and African-American patients in the first year after KAS implementation, without a change in overall pDDKT rate (6, 7). Nevertheless, a longer-term study showed that children < 6 years old were 21% less likely to undergo DDKT after KAS implementation (8). Moreover, a recent report showed that pDDKT recipients < 10 years old experienced a 69% increase in the odds of delayed graft function (DGF) after KAS implementation, and recipients with DGF had a 2.2-fold increase in graft failure compared to those without DGF (9). Finally, another study showed that the overall percentage of pDDKT recipients that received a pediatric donor kidney decreased post-KAS (10). These changes under KAS have prompted some concerns that KAS violates the ethical principles of utility (pediatric recipients are less frequently receiving high-quality kidneys from pediatric donors) and justice (decreased access to transplantation for pediatric candidates < 6 years old), and some have proposed modifications to KAS to attempt to reverse this change (11, 12).

Before considering a change to allocation policy, we felt it was important to have a complete understanding of the effect of KAS on pDDKT recipients. To better inform this discussion, we explored whether KAS-related changes have led to worse post-transplant outcomes in pDDKT. We chose to study patient and graft survival, delayed graft function, and length of stay since these are important post-transplant outcomes for patients, families, and physicians, and could have been affected by KAS-related changes in donor/recipient case-mix. For example, a decrease in pediatric donors being allocated to pediatric recipients might lead to decreased graft survival, and an increase in highly sensitized recipients and recipients with a longer pre-transplant dialysis time might lead to an increased risk of DGF and longer LoS) (9, 10). Therefore, we analyzed national registry data to determine whether post-transplant outcomes (patient and graft survival, delayed graft function, and length of stay) for pDDKT recipients have worsened following KAS implementation. As a result of some of the negative consequences of KAS described above, we hypothesized that some of these post-transplant outcomes may have worsened under KAS.

METHODS

Data source

This study used data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donors, wait-listed candidates, and transplant recipients in the US, submitted by the members of the Organ Procurement and Transplantation Network (OPTN), and has been described elsewhere (13). The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services provides oversight to the activities of the OPTN and SRTR contractors.

Study population

We compared pDDKT (< 18 years old) recipients who were transplanted pre-KAS (12/4/2012 to 12/3/2014) to those transplanted post-KAS (12/4/2014 to 12/3/2016). This study was approved by the Johns Hopkins University Institutional Review Board.

Mortality and graft loss

We analyzed mortality and all-cause graft loss using Cox proportional hazards regression with a sandwich estimator, which accounts for clustering of outcomes by center (14). We first used a univariable model, and then created a multivariable model, adjusting for the following recipient and donor factors: recipient age at transplant, gender, race/ethnicity, ABO blood type, years on dialysis, CPRA at transplant (0–19%, 20–94%, 95–100%), cause of end-stage renal disease (ESRD), donor age, cold ischemia time (CIT) in hours, KDPI (as a binary variable greater or less than 35%), organ share type (local, regional, or national), and receipt of a zero-mismatch organ. To analyze mortality, we administratively censored the pre-KAS cohort on 5/31/2015 and the post-KAS cohort on 5/31/2017, to ensure equal follow-up times between groups. We censored both cohorts at an earlier date for analysis of graft loss, since ascertainment of graft loss in SRTR lags behind that of mortality (pre-KAS cohort on 01/01/2015 and the post-KAS cohort on 01/01/2017).

Trends in graft loss

To determine if our results were due to secular trends in graft loss, we included additional groups of pDDKT recipients who were transplanted from 12/4/2006 to 12/3/2008, 12/4/2008 to 12/3/2010, and 12/4/2010 to 12/3/2012. We administratively censored these groups on 01/01/2009, 01/01/2011, and 01/01/2013, respectively, to ensure equal follow-up time to the pre-KAS and post-KAS cohort. We then included a continuous time variable in our regression model to determine whether there were any secular trends in graft loss during the study period.

Delayed graft function and length of stay

We also examined changes in the incidence of DGF and LoS following KAS implementation. DGF was defined as the need for dialysis within the first week after transplantation, as reported to the OPTN. We analyzed DGF using logistic regression, excluding those who received a pre-emptive transplant (448 patients, or 23.7% of the original cohort). We analyzed LoS using negative binomial regression, excluding patients for whom their index hospitalization LoS was missing (65 patients, or 3.4% of the original cohort). We included a sandwich estimator in both models to account for within center clustering of outcomes (14). Both DGF and LoS were first analyzed with a univariable model, then a multivariable model adjusting for the same donor and recipient characteristics as for graft loss and mortality.

Effect modifiers of graft loss

In our primary analysis, we found that post-KAS recipients had less graft loss than pre-KAS recipients. In order to understand post-KAS changes that may have resulted in this, we explored potential effect modifiers using interaction terms, with a separate unadjusted model for each interaction. Interactions between the donor and recipient characteristics used in our multivariable model were explored.

Sensitivity Analyses

Since prior studies have shown that younger recipients may have been uniquely disadvantaged by KAS, we also separately examined post-transplant outcomes among recipients < 6 years old and < 10 years old as a sensitivity analysis (8, 9). Additionally, we performed a sensitivity analysis for the development of DGF including pre-emptive pDDKT recipients, as they comprised almost 25% of our initial sample population.

Statistical analysis

We used the Wilcoxon rank-sum test to compare baseline differences in age, years on dialysis, CIT, and KDPI. We used Pearson’s Chi-squared test or Fisher’s exact test to compare baseline differences in other categorical variables as appropriate. Data was missing for 2.8% of our study population. We used Little’s test to determine whether our data was missing completely at random (15). This test was non-significant, suggesting that the assumption that our data was missing completely at random was reasonable. Therefore, we performed complete case analysis for our adjusted regressions, where observations with missing data were excluded. Confidence intervals are reported as per the method of Louis and Zeger (16). All analyses were performed using Stata 14.1/MP for Windows (College Station, Texas).

RESULTS

Study Population

Pre-KAS and post-KAS recipients were a similar age (median 13 years vs. 13, p=0.8) and were equally likely to be female (41.9% vs. 43.3%, p=0.5), white (38.4% vs. 40.6%, p=0.4), blood type O, (50.7% vs. 51.0%, p=0.7), have congenital anomalies of kidney and urinary tract as the cause of ESRD (29.3% vs. 29.7%, p=0.6), and receive nationally shared kidneys (4.2% vs. 4.1%, p=0.7) (Table 1). Post-KAS recipients spent more time on dialysis (median 1.26 years vs. 1.07, p=0.02), were more likely to have 100% CPRA (2.0% vs. 0.1%, p=0.001), and were equally like to have ≤ 3 mismatches with their donor (17.1% vs. 16.1%, p=0.5). Donor age increased slightly post-KAS (median [IQR] 22 years [18 – 29] vs. 22 years [17 – 27], p<0.001), and post-KAS recipients were less likely to receive an organ from a pediatric donor (22.6% vs. 28.1%, p<0.01). CIT was not significantly different between groups (median 11.6 hours vs. 11.5 hours, p=0.3), nor was KDPI (median 12 vs. 11, p=0.4).

Table 1.

Characteristics of pre-KAS recipients compared to post-KAS recipients.

	Pre-KAS	Post-KAS	p-value	Missingness (%)
	N=953	N=934
Recipient Characteistics
Age at tx, median (IQR)	13 (7, 16)	13 (7, 16)	0.8	0
Female	399 (41.87%)	404 (43.25%)	0.6	0
Race/ethnicity			0.4	0
White	366 (38.51%)	379 (40.58%)
Hispanic	247 (25.92%)	209 (22.38%)
African-American	38 (3.99%)	46 (4.93%)
Asian	271 (28.33%)	269 (28.80%)
Others	31 (3.25%)	31 (3.32%)
Blood type			0.7	0
A	305 (32.00%)	314 (33.62%)
B	124 (13.01%)	108 (11.56%)
AB	41 (4.30%)	36 (3.85%)
O	483 (50.68%)	476 (50.96%)
Years on dialysis, median (IQR)	1.07 (0.06, 2.27)	1.26 (0.11, 2.63)	0.02	0.2
CPRA at tx			0.001	0
0–19%	789 (82.79%)	777 (83.19%)
20–94%	151 (15.84%)	128 (13.70%)
95–97%	8 (0.84%)	5 (0.54%)
98%	2 (0.21%)	4 (0.43%)
99%	2 (0.21%)	1 (0.11%)
100%	1 (0.10%)	19 (2.03%)
Diagnosis at transplant			0.6	0
CAKUT*	279 (29.28%)	277 (29.66%)
Other familial/metabolic	59 (6.19%)	59 (6.32%)
GN*	46 (4.83%)	32 (3.43%)
Focal Glomerularsclerosis	113 (11.86%)	118 (12.63%)
Other	456 (47.85%)	448 (47.97%)
Donor Characteristics
Age, median (IQR)	22 (17, 27)	22 (18, 29)	<0.001	0
Cold ischemia time in hours, median (IQR)	11.5 (8.0, 15.3)	11.6 (8.0, 16.3)	0.3	2.5
KDPI*, median (IQR)	11 (6, 22)	12 (6, 21)	0.4	0
Sharing			0.7
Local	873 (91.61%)	863 (92.40%)
Regional	40 (4.20%)	33 (3.53%)
National	40 (4.20%)	38 (4.07%)
HLA mismatches ≤ 3	163 (17.1%)	150 (16.1%)	0.5	0.05

^*CAKUT, Congenital Anomalies of Kidney and Urinary Tract; GN, Glomerularonephritis; KDPI, Kidney Donor Profile Index, calculated using 2017 as the reference year.

Graft loss

The cumulative incidence of graft loss at 2-years post-transplantation was 10.2% for pre-KAS recipients and 4.2% for post-KAS recipients, after a median follow-up time of 1.03 years (Figure 1). Through 2-years post-transplant, the unadjusted hazard of graft loss was lower for post-KAS recipients compared to pre-KAS recipients (hazard ratio [HR]: _0.350.54_0.83, p=0.005; Table 2). However, there was no statistically significant post-KAS difference in graft loss for recipients < 10 years old (HR: _0.350.78_1.70, p=0.5), or for recipients < 6 years old (HR: _0.150.44_1.30, p=0.1).

Figure 1. All-cause graft loss by calender year.

Each line represents recipients from five separate groups: recipients from 2006–2008, 2008–2010, 2010 – 2012, the pre-KAS cohort, and the post-KAS cohort. The cumulative incidence of grft loss was lowest in the post-KAS group.

Table 2. Relative Risk of events post-KAS compared to pre-KAS.

Proportional hazard Cox regressions were used for graft loss and patient mortality. Logistic regression was used for delayed graft function. Negative binomial regression was used for length of stay.

	Age<18	Age 10–17	Age<10	Age<6	Age<18 Adjusted
Graft loss (HR)	0.350.540.83	0.250.450.81	0.350.781.70	0.150.441.30	0.380.590.91
Patient mortality (HR)	0.290.721.83	0.030.343.99	0.280.973.37	0.230.853.07	0.371.042.92
DGF (OR)	0.901.321.93	0.871.342.07	0.671.282.45	0.581.302.95	0.941.412.13
LoS (LoS ratio)	0.941.061.19	0.941.091.25	0.831.001.21	0.800.981.20	0.931.041.16

HR, hazard ratio; OR, odds ratio; LoS, length of stay; DGF, delayed graft function

All ratios denote the increased or decreased risk for post-KAS recipients compared to pre-KAS recipients. Bolded values represent a ratio that significantly different than 1.0 (p<0.05). DGF results are with pre-emptive recipients excluded, although inferences remain the same with these recipients included.

In order to determine whether there were secular changes in graft loss preceeding KAS that could account for the post-KAS improvement in graft loss, we examined graft loss in recipients from successive two-year time periods preceeding the pre-KAS group (2006–2008, 2008–2010, and 2010–2012). The cumulative incidence of graft loss at 2-years post-transplant was 11.2% for 2006–2008 recipients, 10.5% for 2008–2010 recipients, and 6.1% for 2010–2012 recipients (Figure 1). In our model adjusted for a continuous trend over time, there was no relationship between time and risk of graft loss (p=0.9), suggesting that there were no underlying secular trends responsible for the improved graft loss post-KAS.

In order to understand potential causes for this apparent decrease in graft loss in post-KAS recipients, we adjusted for donor and recipients characteristics, but this did not appreciably change our results (adjusted HR [aHR]: _0.380.59_0.91, p=0.02; Table 2). None of the potential interaction terms between KAS and recipient and donor characteristics were significant.

Mortality

The cumulative incidence of mortality at 2-years post-transplantation was 1.1% for pre-KAS recipients and 0.6% for post-KAS recipients, after a median follow-up time of 1.47 years. Through 2-years post-transplantation, the unadjusted hazard of mortality was not significantly different between post-KAS recipients and pre-KAS recipients (HR: _0.290.72_1.83, p=0.6; Table 2). This remained consistent after adjusting for donor and recipient characteristics (aHR: _0.371.04_2.92, p=0.9; Table 2). There were no statistically significant post-KAS differences in mortality for recipients < 10 years old (HR: _0.280.97_3.37, p=1.0) or < 6 years old (HR: _0.230.85_3.07, p=0.8; Table 2).

Delayed Graft Function

Among patients who were not transplanted pre-emptively, 8.7% of pre-KAS recipients developed DGF compared to 11.1% of post-KAS recipients. Post-KAS recipients were not significantly more likely to develop DGF compared to pre-KAS recipients (odds ratio[OR]: _0.901.32_1.93, p=0.2, Table 2). This remained consistent after adjusting for donor and recipient characteristics (adjusted OR: _0.941.41_2.13, p=0.1; Table 2). There were no statistically significant differences in development of delayed graft function in recipients < 10 years old (OR: _0.671.28_2.45, p=0.5) or for recipients < 6 years old (OR: _0.581.30_2.95, p = 0.5). In our sensitivity analysis including pre-emptive pDDKT recipients, 6.9% of pre-KAS recipients developed DGF compared to 9.4% of all post-KAS recipients. These recipients were not significantly more likely to develop DGF post-KAS compared to pre-KAS (_0.951.40_2.06, p=0.1). There were no statistically significant differences after adjusting for donor and recipient characteristics (adjusted OR: _0.981.49_2.25, p=0.06).

Length of Stay

The median (IQR) length of stay for pre-KAS recipients was 8 days (6 – 11), and for post-KAS recipient it was 8 days (6 – 12). LoS for post-KAS recipients was not significantly different compared to pre-KAS recipients (LoS ratio: _0.941.06_1.19, p=0.4, Table 2). This remained consistent after adjusting for donor and recipient characteristics (adjusted LoS ratio: _0.931.04_1.16, p=0.5; Table 2). There were no statistically significant differences in LoS for recipients < 10 years old (LoS ratio: _0.831.00_1.21, p=1.0) or for recipients < 6 years old (LoS ratio: _0.800.98_1.20, p=0.8).

DISCUSSION

In this national study of pDDKT recipients, we did not find any evidence that post-transplant mortality or LoS worsened in the first two years post-KAS. Although there was no statistically significant change in the incidence of DGF, our point estimate approached statistical significance despite a relatively small sample size. We also showed that graft loss decreased post-KAS, although this was not explained by any KAS-related changes in donor or recipient characteristics that we studied. Additionally, there were no underlying secular trends in graft loss during the study period that could account this improvement. We also did not observe any statistically significant worsening of outcomes in younger recipients, who have been shown to be specifically affected by KAS (8, 9).

The implementation of KAS has led to many positive changes for the adult DDKT population. Sensitized patients, racial minorities, and certain blood types now have improved access to DDKT (6, 7, 17–23). Nevertheless, one study in the pDDKT population showed that recipients are less likely to receive a pediatric kidney after KAS as a consequence of using KDPI to allocate kidneys, an index which does not fully capture the unique considerations of pediatric donors (low height and weight, characteristics such as en bloc vs. single graft) and may not accurately predict graft survival for pediatric recipients (10, 24). The KDPI formula assigns higher scores to kidneys from younger, and consequentially smaller, donors. Therefore, kidneys from pediatric donors may be assigned a high KDPI that precludes allocation to pediatric recipients (24). Although a decrease in allocation of pediatric kidneys to pDDKT recipients is concerning, our results suggest that this change has not led to worsened post-transplant graft survival. In contrast, we have shown that graft survival was actually improved in the two years post-KAS compared to the two years pre-KAS, in the absence of any secular trends. One potential explanation for improved graft loss might be improved recipient selection, through mechanisms not currently measured in registry data. Additionally, an underlying change in management of pediatric recipients, unrelated to KAS, might also explain the improved two-year graft survival. However, to our knowledge there have been no recent large-scale changes in the management of pDDKT recipients whose effect would be stronger than the effect of KAS, which represents the largest change to the allocation system in almost two decades.

We also did not find any statistically significant changes in the incidence of DGF and LoS following pDDKT after KAS implementation. DGF is an important outcome to study since it is associated with worse post-transplant outcomes, including acute rejection and graft failure (25–28). Moreover, it is a relatively more common complication after DDKT, such that any differences in its incidence after KAS should be readily apparent. Our finding of no statistically significant difference in DGF rates after KAS is in contrast to a study that reported a 69% increase in the odds of DGF for pDDKT recipients < 10 years old (8). This study included five years of pre-KAS recipients and two years of post-KAS recipients, focusing mainly on recipients < 10 years old. In contrast, we limited our pre-KAS cohort to only include two years of all pDDKT recipients, in order to increase its comparibility to our post-KAS cohort and limit the impact of any secular trends. As such, their study was more powered than ours was for the analysis of recipients < 10 years old, and might explain the difference in our findings. We did not find a statistically significant difference, in either overall or subgroup specific rates of DGF. Given our small sample size, it is possible that we were underpowered to detect a true difference of this size. However, in the context of an increase in DGF reported in the adult literature, the prior pediatric study suggesting an increased risk of DGF for recipients < 10 years old, and our results which approached statistical significance despite our small sample size, there may be a true increase in DGF rates for pDDKT recipients post-KAS that we are underpowered to detect in the current study. We also did not find a change in LoS for post-KAS recipients, which is consistent with a report from the adult literature (29). This is reassuring given that the mean pre-transplant dialysis time and the proportion of cPRA 100% recipients increased significantly in our study post-KAS, both of which are associated with higher rates of DGF and could potentially have increased LoS (27, 28).

It has recently been suggested that KAS, as it stands now for pediatric patients, violates fundamental ethical principles, and that KAS should be modified to reverse some of the changes it instituted (11, 12). This has been based on arguments for decreased justice (decreased access to transplantation for children < 6 years old) and utility (the highest quality pediatric donor kidneys are not being as frequently directed to children, who have the longest expected post-transplant survival) (12). Without question, critical appraisal of KAS is important, and an ongoing evaluation of how it is affecting pediatric candidates is necessary to ensure that pediatric candidates continue to receive relative prioritzation.

Our results add context to this discussion, and underscore that KAS has had a complex effect on the pediatric population. It has led to decreased DDKT rates for candidates < 6 years, a potentially increased incidence of DGF, but also decreased graft loss. As the transplant community grapples with whether, or how, to modify KAS for pediatric patients, our results add an understanding of how KAS has affected a range of post-transplant outcomes to these discussions. Although pre-transplant outcomes of pediatric candidates are outside of the scope of the current study (i.e. equity in access to DDKT), we felt it was important to study post-transplant outcomes of pediatric recipients, who are equally important stakeholders in, and equally affected by KAS. Ultimately, discussions of KAS modifications should consider the entire context of both pre-transplant and post-transplant changes for pediatric candidates.

Our study has several noteworthy limitations. In using national registry data, we depend on accurate outcome ascertainment. In pDDKT recipients, graft loss is a rare event and therefore inaccurate graft loss ascertainment could bias our results. However, we standardized follow-up times between pre-KAS and post-KAS recipients to minimize differences in ascertainment across the two eras. We are also limited to only two years of follow-up after KAS implementation. The changes in outcomes (or lack thereof) we have demonstrated here may change over time, especially for longer-term outcomes such as patient and graft survival. Despite this, we feel it is important to document short-term trends, to identify any unintended consequences of KAS. Additionally, as with any observational study using national registry data, we are unable to account for factors not captured by SRTR that might influence post-transplant outcomes (i.e. unmeasured confounding). Finally, studies of pediatric transplant recipients are limited by small sample sizes, and lower statistical power, relative to studies of adult recipients, and our study is no exception. Although we did not find any statistically significant post-KAS differences in mortality or DGF, this does not necessarily mean that there have not been smaller changes (but clinically important) that we are underpowered to detect. Although there is no analytical technique to overcome this weakness, our results should be considered in the full context of our sample size, point estimates, confidence intervals, and p-values.

In conclusion, our results suggest that KAS has had a mixed effect on post-transplant outcomes for pDDKT recipients, although they should be considered in the context of a relatively small sample size and short-term follow-up. There were no statistically significant changes in mortality, LoS, or DGF. The small increase in the post-KAS incidence of DGF we report approached statistical significance, and in the context of other adult and pediatric literature, may represent a true increase ing DGF post-KAS. Conversely, two-year graft survival was significantly improved post-KAS, although this was not adequately explained by changes in donor or recipient characteristics that we studied. Nevertheless, ongoing assessment of post-transplant outcomes should continue, to ensure that the highest-quality data is being used to critically evaluate the ongoing effects of KAS.

Abbreviations:

CIT

Cold Ischemia Time

DDKT

Deceased Donor Kidney Transplantation

pDDKT

Pediatric Deceased Donor Kidney Transplantation

DGF

Delayed Graft Function

OPTN

Organ Procurement and Transplantation Network

KAS

Kidney Allocation System

SRTR

Scientific Registry for Transplant Recipients

KDPI

Kidney Donor Profile Index

KDRI

Kidney Donor Risk Index

IQR

Interquartile Range

HR

Hazard Ratio

OR

Odds Ratio

aOR

adjusted odds ratio

LoS

Length of Stay