Genetic predisposition to acute kidney injury in critically ill adults

Division of Intensive Care Medicine

Department of Anesthesiology, Intensive Care and Pain Medicine University of Helsinki and Helsinki University Hospital

Doctoral Programme in Clinical Research Faculty of Medicine

University of Helsinki

'ĞŶĞƟĐƉƌĞĚŝƐƉŽƐŝƟŽŶƚŽĂĐƵƚĞŬŝĚŶĞLJŝŶũƵƌLJ ŝŶĐƌŝƟĐĂůůLJŝůůĂĚƵůƚƐ

Laura M. Vilander

ACADEMIC DISSERTATION

To be presented with the permission of the Faculty of Medicine of the University of Helsinki, for public examination in Lecture Hall 2, Haartman Institute,

on December 13^th 2019, at 12 noon.

Helsinki 2019

Supervisors

Professor Ville Pettilä

Department of Anesthesiology, Intensive Care and Pain Medicine,

University of Helsinki and Helsinki University Hospital, Helsinki, Finland

Docent Mari Kaunisto

Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki,

Helsinki, Finland Reviewers

Docent Satu Mäkelä

Department of Internal Medicine, Tampere University,

Tampere, Finland Docent Tarja Kunnas

Department of Medical Biochemistry, Tampere University,

Tampere, Finland

ˆϐ‹…‹ƒŽ’’‘‡–

Professor Jean-Daniel Chiche

Department of Critical Care Medicine, Université Paris Descartes,

Paris, France

The Faculty of Medicine uses the Urkund system (plagiarism recognition) to examine all doctoral dissertations.

ISBN 978-951-51-5530-6 (paperback) ISBN 978-951-51-5531-3 (PDF)

To My Family

Table of contents

LIST OF ORIGINAL PUBLICATIONS ... 6

LIST OF ABBREVIATIONS ... 7

ABSTRACT ... 10

1 INTRODUCTION ... 12

2 REVIEW OF THE LITERATURE ... 13

2.1 HUMAN GENETICS ... 13

Ϯ͘ϭ͘ϭKǀĞƌǀŝĞǁŽĨƚŚĞƐƚƌƵĐƚƵƌĞĂŶĚĨƵŶĐƟŽŶŽĨƚŚĞŚƵŵĂŶŐĞŶŽŵĞ ... 13

Ϯ͘ϭ͘ϮsĂƌŝĂƟŽŶŝŶƚŚĞŚƵŵĂŶŐĞŶŽŵĞ ... 14

Ϯ͘ϭ͘ϯ/ŵƉĂĐƚŽĨŐĞŶĞƟĐǀĂƌŝĂƟŽŶ ... 16

2.2 IDENTIFYING SUSCEPTIBILITY VARIANTS FOR COMPLEX DISEASES ... 18

2.2.1 Heritability ... 18

Ϯ͘Ϯ͘ϮŽŵƉůĞdžĚŝƐĞĂƐĞƐĂŶĚƚƌĂŝƚƐ ... 18

Ϯ͘Ϯ͘ϯ'ĞŶĞƟĐƐƚƵĚLJĚĞƐŝŐŶƐ ... 20

Ϯ͘Ϯ͘ϯ͘ϭ>ŝŶŬĂŐĞƐƚƵĚŝĞƐ ... 20

Ϯ͘Ϯ͘ϯ͘ϮƐƐŽĐŝĂƟŽŶƐƚƵĚŝĞƐ ... 20

Ϯ͘Ϯ͘ϯ͘ϯEŽǀĞůƐĞƋƵĞŶĐŝŶŐƚĞĐŚŶŽůŽŐŝĞƐ ... 21

2.3 ACUTE KIDNEY INJURY (AKI) ... 21

Ϯ͘ϯ͘ϭZĞŶĂůĨƵŶĐƟŽŶ ... 21

Ϯ͘ϯ͘Ϯ</ĚĞĮŶŝƟŽŶ ... 22

Ϯ͘ϯ͘ϯWĂƚŚŽƉŚLJƐŝŽůŽŐLJŽĨ</ ... 23

2.3.4 Risk factors for AKI ... 25

Ϯ͘ϯ͘ϱ</ĞƉŝĚĞŵŝŽůŽŐLJ ... 26

Ϯ͘ϯ͘ϲ</ƚƌĞĂƚŵĞŶƚŵŽĚĂůŝƟĞƐ ... 27

3 AIMS OF THE STUDY ... 28

4 MATERIALS AND METHODS ... 29

4.1 PATIENTS ... 29

ϰ͘ϭ͘ϭŝĂŐŶŽƐƟĐĚĞĮŶŝƟŽŶƐ ... 31

4.1.1.1 Acute kidney injury ... 31

4.1.1.2 Chronic kidney disease (CKD) ... 32

ϰ͘ϭ͘ϭ͘ϯ^ĞƉƐŝƐ ... 32

4.2 DATA COLLECTION ... 33

4.3 STUDY DESIGNS ... 34

4.3.1 Study I ... 34

4.3.2 Study II ... 34

4.5 DNA PROCESSING AND GENOTYPING ... 37

ϰ͘ϱ͘ϭEĞdžƚƌĂĐƟŽŶĂŶĚƐĂŵƉůĞƉƌĞƉĂƌĂƟŽŶ ... 37

ϰ͘ϱ͘ϮsĂƌŝĂŶƚƐĞůĞĐƟŽŶ ... 38

ϰ͘ϱ͘ϯdĂƌŐĞƚĞĚƐŝŶŐůĞŶƵĐůĞŽƟĚĞƉŽůLJŵŽƌƉŚŝƐŵŐĞŶŽƚLJƉŝŶŐ ... 40

ϰ͘ϱ͘ϰ&ƌĂŐŵĞŶƚĂŶĂůLJƐŝƐ ... 41

4.5.5 Quality control ... 41

ϰ͘ϱ͘ϱ͘ϭ,ĂƌĚLJͲtĞŝŶďĞƌŐĞƋƵŝůŝďƌŝƵŵ;,tͿ ... 42

4.6 ENZYME CONCENTRATION ANALYSIS ... 42

4.7 STATISTICAL METHODS ... 43

4.8 ETHICAL ASPECTS OF THE STUDY ... 46

5 RESULTS ... 47

5.1 QUALITY OF AKI ASSOCIATION STUDIES (I) ... 47

5.2 GENETIC VARIANTS ASSOCIATED TO AKI ACCORDING TO SYSTEMATIC REVIEW (I) ... 48

5.3 REPLICATION OF PREVIOUS ASSOCIATION FINDINGS ... 50

ϱ͘ϯ͘ϭƉŽƉƚŽƐŝƐͲƌĞůĂƚĞĚǀĂƌŝĂŶƚƐ;//Ϳ ... 52

ϱ͘ϯ͘Ϯ/ƌŽŶŵĞƚĂďŽůŝƐŵͲƌĞůĂƚĞĚǀĂƌŝĂŶƚ;///Ϳ ... 54

ϱ͘ϯ͘ϯ/ŶŇĂŵŵĂƟŽŶͲƌĞůĂƚĞĚǀĂƌŝĂŶƚƐ;/sͿ ... 55

6 DISCUSSION ... 62

6.1 GENETIC ASSOCIATION STUDIES REGARDING AKI (I) ... 62

6.2 APOPTOSIS RELATED GENETIC VARIATION IN AKI (II) ... 64

6.3 IRON METABOLISM RELATED GENETIC VARIATION IN AKI (III) ... 66

6.4 INFLAMMATION-RELATED GENETIC VARIATION IN AKI (IV) ... 68

6.5 LIMITATIONS ... 70

6.6 CLINICAL IMPLICATIONS ... 72

6.7 FUTURE PERSPECTIVES ... 73

7 CONCLUSIONS ... 76

8 ACKNOWLEDGEMENTS ... 77

9 REFERENCES ... 79

LIST OF ORIGINAL PUBLICATIONS

This thesis is based on the following original articles, which are referred to in the text by their Roman numerals (I-IV):

I Vilander LM, Kaunisto MA, Pettilä V. Genetic predisposition to acute kidney injury – a systematic review. BMC Nephrol.

2015;16:197.

II Vilander LM, Kaunisto MA, Vaara ST, Pettilä V; FINNAKI Study Group. Genetic variants in SERPINA4 and SERPINA5, but not BCL2 and SIK3 are associated with acute kidney injury in critically ill patients with septic shock. Crit Care. 2017;21:47.

III Vilander LM, Vaara ST, Donner KM, Lakkisto P, Kaunisto MA, Pettilä V; FINNAKI Study Group. Heme Oxygenase-1 Repeat Polymorphism In Septic Acute Kidney Injury. PLoS One.

2019; 14(5): e0217291.

IV Vilander LM, Kaunisto MA, Vaara ST, Pettilä V; FINNAKI Study

”‘—’Ǥ‘‘‹ϐŽƒƒ–‹‘Ǧ”‡Žƒ–‡†…ƒ†‹†ƒ–‡‰‡‡˜ƒ”‹ƒ–•

and acute kidney injury in 2647 critically ill Finnish patients.

J Clin Med. 2019;8(3):342.

LIST OF ABBREVIATIONS

A Adenine

ACCP/SCCM American College of Chest Physicians/Society of Critical Care Medicine

ACE Angiotensin-converting enzyme

AGT Angiotensinogen

ADQI Acute Dialysis Quality Initiative AKI Acute kidney injury

AKIN Acute Kidney Injury Network APC Activated protein C

APACHE Acute Physiology and Chronic Health Evaluation

APOE Apolipoprotein E

ARB Angiotensin II receptor blocker ARDS Acute respiratory distress syndrome ARF Acute renal failure

ATP Adenosine triphosphate BBS9 Bardet-Biedl syndrome 9 BCL2 B-cell CLL/lymphoma 2

BMI Body mass index

C Cytosine

‘ϐ‹†‡…‡‹–‡”˜ƒŽ

CI-AKI Contrast induced acute kidney injury CKD Chronic kidney disease

CO Carbon monoxide

COMT Catechol-O-methyltransferase COPD Chronic obstructive pulmonary disease

CPB Cardiopulmonary bypass

CRF Case report form

CSA-AKI Cardiac surgery-associated acute kidney injury

‘‡ˆϐ‹…‹‡–‘ˆ˜ƒ”‹ƒ–‹‘

CXCL8 Interleukin 8 (gene)

CYBA Cytochrome b245 alpha subunit

DNA Deoxyribonucleic acid

ELISA Enzyme-linked immunosorbent assay EPO Erythropoietin

FCGR2A Receptor IIa of the Fc portion of immunoglobulin G

FICC Finnish Intensive Care Consortium FINNAKI Finnish Acute Kidney Injury G Guanine

Ž‘‡”—Žƒ”ϐ‹Ž–”ƒ–‹‘”ƒ–‡

GWAS Genome wide association study HAMP Hepcidin antimicrobial peptide HDL High density lipoprotein

HIF1A Hypoxia-inducible factor 1-alpha HLA-DRB Human leukocyte antigen-DR-beta HMOX Heme ogygenase-1 (gene)

HO-1 Heme oxygenase-1

HRQL Health-related quality-of-life Hsp27 Heat shock protein 27

HSPB1 Heat shock protein family B (small) member 1 HuGENet Human Genome Epidemiology Network HWE Hardy-Weinberg equilibrium

ICU Intensive Care Unit

IL6 Interleukin 6

IL8 Interleukin 8

IL10 Interleukin 10

IQR Interquartile range

KDIGO Kidney Disease: Improving Global Outcomes

LCT Lactase (gene)

LD Linkage disequilibrium LOD Logarithm of the odds

MALDI-TOF MS Matrix-Assisted Laser Desorption/Ionization Time-Of- Flight Mass Spectrometry

‘†‹ϐ‹…ƒ–‹‘‘ˆ‹‡–‹‡ƒŽ‹•‡ƒ•‡

MPO Myeloperoxidase

NADPH Nicotinamide adenosine dinucleotide phosphate NFKB1A Nuclear factor of kappa light polypeptide gene

enhancer in B-cells inhibitor, alpha

NO Nitric oxide

NOS3 Nitric oxide synthase 3 gene

RIFLE Risk, Injury, Failure, Loss, End-stage

RNA Ribonucleic acid

RRT Renal replacement therapy

SA-AKI Sepsis associated acute kidney injury

‹’Ž‹ϐ‹‡†…—–‡Š›•‹‘Ž‘‰›…‘”‡

SERPINA4/5 Serpin peptidase inhibitor, clade A (alpha-1 antiproteinase, antitrypsin), member 4/5 SFTPD Surfactant protein D (gene)

SIK3 Salt-inducible kinase family 3

›•–‡‹…‹ϐŽƒƒ–‘”›”‡•’‘•‡•›†”‘‡

SNP Single nucleotide polymorphism SOFA Sequential Organ Failure Assessment SUFU Suppressor of fused homolog

T Thymine

Ƚ —‘”‡…”‘•‹•ˆƒ…–‘”ƒŽ’Šƒ

TNFA Tumor necrosis factor alpha (gene) VEGF Vascular endothelial growth factor

VEGFA Vascular endothelial growth factor A (gene)

ABSTRACT

KďũĞĐƟǀĞƐ

Acute kidney injury (AKI) is a complex syndrome that causes increased mortality and morbidity, especially in the critically ill. As clinical factors explain only part of the risk for AKI, individual susceptibility through genetic variation should be considered. The aims of this study were to systematically review the current literature for genetic predisposition to ǡƒ†–‘”‡’Ž‹…ƒ–‡’”‡˜‹‘—•ϐ‹†‹‰•‹‰‡‡•ƒ••‘…‹ƒ–‹‰™‹–Šƒ’‘’–‘•‹•ǡ

‹”‘‡–ƒ„‘Ž‹•ǡƒ†‹ϐŽƒƒ–‹‘‹…”‹–‹…ƒŽŽ›‹ŽŽƒ†—Ž–•Ǥ

Methods

Study I was a systematic review about genetic predisposition to AKI risk and AKI-related outcomes. The review included 28 original studies. Studies II to IV included patients from the prospective, observational FINNAKI study that was conducted in 17 Finnish intensive care units (ICUs) in 2011 and 2012. Emergency admissions and elective admissions with an expected stay longer than 24h were included.

Study II investigated the development of severe AKI and the association between variants in apoptosis-related genes: BCL2; SERPINA4; SERPINA5;

and SIK3. Altogether 478 patients with septic shock were included. In Study III, the association between dinucleotide repeats in the intron of HMOX1 gene and the development of severe AKI was studied in 653 septic patients. In Study IV, the association between 27 candidate polymorphisms and the development of AKI was studied. The tested genetic variants were located within genes that have been previously linked with AKI.

Results

In Study I, the quality of the original studies was evaluated, but no meta- analysis was performed because of the heterogeneity of the studies. The studies gave no conclusive evidence. The majority of the variants were

Ž‘…ƒ–‡† ™‹–Š‹ ‹ϐŽƒƒ–‘”› ‰‡‡• ƒ† ˜ƒ•‘‘–‘”Ǧ”‡‰—Žƒ–‹‘Ǧ”‡Žƒ–‡†

genes with AKI development and AKI related outcomes. One investigation

™‹–ŠƒŠ›’‘–Š‡•‹•Ǧˆ”‡‡•–—†›†‡•‹‰™ƒ•‹†‡–‹ϐ‹‡†‹–—†›Ǥ–—†›ǡ–Š‡

‘‡‘ˆ–Š‡˜ƒ”‹ƒ–•™‡”‡•‹‰‹ϐ‹…ƒ–Ž›ƒ••‘…‹ƒ–‡†™‹–Š‹ƒ›‘ˆ–Š‡

analyses.

Main conclusions

The systemic review conducted provides no conclusive evidence about the genetic predisposition to AKI. The reviewed studies were of inadequate quality and heterogeneous, omitting the possibility for a meta-analysis. In the phenotype of AKI with septic shock, the apoptosis-related genes are

‹ϐŽ‹…–‡† ‹ ’ƒ–Š‘’Š›•‹‘Ž‘‰›Ǥ ‘”‡‘˜‡”ǡ ‹ …”‹–‹…ƒŽŽ› ‹ŽŽ ’ƒ–‹‡–• ™‹–Š

sepsis the repeat polymorphism in the HMOX1 gene promoter sequence had the opposite risk allele for the development of severe AKI than previously found in cardiac-surgery patients.

In conclusion, the majority of the previously suggested candidate gene

˜ƒ”‹ƒ–•–‡•–‡†‹ƒ••‘…‹ƒ–‹‘™‹–Š†‡˜‡Ž‘’‡–‘ˆ™‡”‡‘–…‘ϐ‹”‡†

in this large multicenter prospective study in critically ill patients.

Keywords

Acute Kidney Injury, Genetic Predisposition, Genetic Variation, Apoptosis,

ϐŽƒƒ–‹‘

1 INTRODUCTION

Acute kidney injury (AKI) is a syndrome that manifests with the

ƒ……——Žƒ–‹‘‘ˆ™ƒ•–‡’”‘†—…–•ǡƒ•™‡ŽŽƒ•†‹•–—”„ƒ…‡•‹ϐŽ—‹†ǡ‡Ž‡…–”‘Ž›–‡ǡ

ƒ† ƒ…‹†Ǧ„ƒ•‡ „ƒŽƒ…‡Ǥ Š‡ †‡‰”‡‡ ‘ˆ †‹ˆϐ‹…—Ž–› ‹• †‡•…”‹„‡† „› Ž‘•• ‘ˆ

normal kidney function.¹

Despite the vigorous search for novel markers of kidney injury,^2–4 the examination of kidney function — and its decline — are currently described by plasma or serum creatinine and urine output.¹ These measurements

•‡”˜‡ƒ••—””‘‰ƒ–‡•ˆ‘”–Š‡‡˜ƒŽ—ƒ–‹‘‘ˆ–Š‡ϐ‹Ž–”ƒ–‹‘”ƒ–‡‘ˆ–Š‡‰Ž‘‡”—Ž‹

‹–Š‡‹†‡›Ǥ˜‹†‡–Ž›ǡ–Š‡•‡†‹ƒ‰‘•–‹…•ƒ”‡Žƒ–‡–‘”‡ϐŽ‡…––Š‡‹‹–‹ƒŽnoxa to the still-functioning kidney.⁵

No single pathophysiology is known to account for AKI.^6,7 Multiple factors contribute to this complex syndrome. Moreover, the understanding of all-cause AKI necessitates phenotypical clustering of differing sub phenotypes according to their etiology and trajectory of the course of the syndrome.^5,7

An individualized approach to the susceptibility to complex diseases has been successfully investigated.^8,9 Genetic variation in association with AKI possibly explains part of the risk. The relevance in searching for individual predisposition lies in more comprehensive understanding of the etiology of AKI. This could ultimately lead to the development of rationally targeted diagnostic, preventive, and therapeutic interventions. In addition, the

‘™Ž‡†‰‡‘ˆ’”‡†‹•’‘•‹‰ˆƒ…–‘”•ƒ••‹•–•‹”‹•’”‘ϐ‹Ž‹‰‘…‡Šƒ”‡••‡†

into a well-performing risk-prediction model.¹⁰ Polygenic risk prediction scores are researched in association to common diseases and are likely to be of clinical use in time.¹¹

AKI is common in all hospitalized patients,^12,13 but in association with critical illness the incidence is substantial.^14,15 The majority of critically ill patients with sepsis are diagnosed with AKI.¹⁶ AKI may result in the permanent loss of kidney function and the permanent need for renal replacement therapy. AKI in the intensive care unit (ICU) is associated with increased mortality.¹⁴ The individual susceptibility to AKI is incompletely understood.

The aim of this study was to systematically review the current literature regarding genetic predisposition to AKI in critically ill adults and, in light of this knowledge, search for genetic variants associated with AKI in a Finnish ICU cohort.

1,

2 REVIEW OF THE LITERATURE

2.1 HUMAN GENETICS

In human genetics, variation within the human genome and the related consequences on human features are researched. Human genetics encompasses, but is not limited to, several overlapping disciplines, such as genomics and medical genetics, as well as population and developmental genetics. In these disciplines, the role of genes is discussed differently, but with a common objective to understand human life. In the following overview, the viewpoint is mostly molecular genetics.

Ϯ͘ϭ͘ϭKǀĞƌǀŝĞǁŽĨƚŚĞƐƚƌƵĐƚƵƌĞĂŶĚĨƵŶĐƟŽŶŽĨƚŚĞŚƵŵĂŶŐĞŶŽŵĞ The euchromatic portion of the human genome sequence was deciphered by the Human Genome Project that concluded its work in 2003.¹⁷ The

Š—ƒ‰‡‘‡‡…‘†‡•‘Ž›ʹͲԜͲͲͲ–‘ʹͷԜͲͲͲ’”‘–‡‹Ǧ…‘†‹‰‰‡‡•¹⁷ that are sparsely distributed.¹⁸ The remainder of the genome is noncoding, with ample highly-repetitive sequences of incompletely known function.¹⁸

The human genome consists of nuclear deoxyribonucleic acid (DNA) within the nucleus and mitochondrial DNA within the mitochondria. In the nucleus, the DNA molecules form 23 pairs of chromosomes in human somatic cells.¹⁹ These pairs share the locations of genes, however, each allele may be different. The individual setup of alleles is the genotype. An individual can be homozygous or heterozygous at a given locus, depending on whether the alleles are the same or if they differ, respectively. This holds true for the 22 pairs of autosomes, but the sex determining allosomes come in two forms, X and Y, that are not complementary.

The DNA molecule is a large polymer that has sugar-phosphate backbone and a varying base that attaches to the sugar, thus forming a strand of nucleotides.¹⁸ The bases of DNA are adenine (A), guanine (G), cytosine (C), and thymine (T). Base pairing creates the DNA duplex form, in which the two DNA strands curve around one another to a double helix around the bases. The bases on opposing strands link together by hydrogen bonding according to a standard combination of the tolerated base pairs: A goes with T, and G with C.¹⁸

The DNA nucleotide sequence is an instruction to generate a complementary ribonucleic acid (RNA) sequence in a process called transcription. A gene is a set of such instructions, with an exon segment that is expressed and intron segments that are not expressed. In addition, in the vast majority of intergenic DNA that remains unexpressed, there are promoter regions to genes that contribute in the gene expression regulation. Even if each human cell contains every gene, the gene

expression needs to be regulated to match the temporal and tissue-

•’‡…‹ϐ‹…‡‡†•Ǥ¹⁸

Š‡ ϐ‹ƒŽ ’”‘†—…– ‘ˆ –”ƒ•…”‹’–‹‘ ‹• ‘†‹ϐ‹‡† „›ǡ ˆ‘” ‡šƒ’Ž‡ǡ splicing and, once mature, will serve as either noncoding operator or a messenger template to make a polypeptide. In a process called translation, the sequence delivered by the messenger is deciphered and an amino- acid chain is created according to nucleotide triplets. Not every triplet corresponds to a unique amino acid and hence not all variation leads to …Šƒ‰‡‹–Š‡‡š’”‡••‡†’‘Ž›’‡’–‹†‡Ǥ‹–Šˆ—”–Š‡”‘†‹ϐ‹…ƒ–‹‘ǡƒ’”‘–‡‹

is generated — an end product of the central dogma of molecular biology.²⁰

Ϯ͘ϭ͘ϮsĂƌŝĂƟŽŶŝŶƚŚĞŚƵŵĂŶŐĞŶŽŵĞ

Variation in the genome occurs because of the inheritance of one allele from each parent, and additionally because of genomic mutations.²¹ Common DNA variants with a population frequency more than 1% are traditionally called polymorphisms.²² ‘™‡˜‡”ǡ ƒ ‘”‡ ˆ‡ƒ•‹„Ž‡ …Žƒ••‹ϐ‹…ƒ–‹‘ ‹•ǣ common variant with more than 5%; low variant between 0.5 and 5%; and rare variant with below 0.5% frequency.²²

In meiosis, the homologous chromosomes exchange genetic material in corresponding loci through recombination, before the random assignment of either of the chromosomes in to a gamete. This increases the inherited variation. However, genetic loci of close proximity are inherited together more often than is expected by chance, since the crossing over does not usually separate them. The alleles in such loci are said to be in linkage disequilibrium (LD). However, LD is affected by other things besides physical distance between the alleles. Stable blocks of alleles in LD were ϐ‹”•– †‡•…”‹„‡† ‹ –Š‡ ƒ’ƒ’ ”‘Œ‡…–Ǥ²³ The region adjacent harbors several such alleles that are known to descend as a unit called a “haplotype”.

Figure 1 illustrates an example of a haploblock in an LD plot.

2

&ŝŐƵƌĞϭ͘ŶŽƟŽŶĂůŝůůƵƐƚƌĂƟŽŶŽĨĂůŝŶŬĂŐĞĚŝƐĞƋƵŝůŝďƌŝƵŵ;>ͿƉůŽƚǁŝƚŚŚĂƉůŽďůŽĐŬ ŽĨǀĂƌŝĂŶƚƐϭƚŽϱ͘^ƚLJůĞĂĚĂƉƚĞĚĨƌŽŵƚŚĂƚƵƐĞĚŝŶ,ĂƉůŽǀŝĞǁƐŽŌǁĂƌĞ͘²⁴dŚĞƌĞĚ ĐŽůŽƌŝŶĚŝĐĂƚĞƐĂŚŝŐŚůĞǀĞůŽĨůŝŶŬĂŐĞ͘EƵŵďĞƌƐŝŶƐŝĚĞƐƋƵĂƌĞƐŝŶĚŝĐĂƚĞŶŽƌŵĂůŝnjĞĚ ĐŽĞĸĐŝĞŶƚŽĨůŝŶŬĂŐĞĚŝƐĞƋƵŝůŝďƌŝƵŵ;͛Ϳ͕ĂĐŚŝĞǀŝŶŐǀĂůƵĞƐďĞƚǁĞĞŶϬĂŶĚϭϬϬ;ŶŽƚ ƐŚŽǁŶͿ͘^ŝnjĞŽĨƚŚĞŚĂƉůŽďůŽĐŬŝƐƐƚĂƚĞĚŝŶŬŝůŽďĂƐĞƐ;ŬďͿ͘

Mutations can occur spontaneously or due to an external trigger. The spontaneous endogenous mutations, such as errors in cellular mechanisms or chemical damage to DNA, exceed the mutations caused by environmental chemicals and radiation.²² The majority of the DNA changes are small scale and rarely have an obvious effect on the phenotype.²² Structural variation is less frequent, but usually has large consequences.²² The borderline between these, however, is arbitrary. In Table 1 some common types of genetic variation are listed according to the nomenclature from most recent recommendation.²⁵

Block 1 (N kb)

1 2 3 4 5 6

98 0

99 99 99 0 98 98 2

98 0 0

Variant 1 Variant 2 Variant 3 Variant 4 Variant 5 Variant 6

Ϯ͘ϭ͘ϯ/ŵƉĂĐƚŽĨŐĞŶĞƟĐǀĂƌŝĂƟŽŶ

The mere knowledge of the position of genetic variation provides little understanding in relation to its impact. There are several databases that contain information about the predicted functional consequence of genetic variants (ClinVar²⁶, Ensembl²⁷). In addition, biochemical and cellular assays are the most reliable proof of the function of a genetic variant.

Table 1.ŽŵŵŽŶƚLJƉĞƐŽĨŐĞŶĞƟĐǀĂƌŝĂƟŽŶ͘

dLJƉĞŽĨǀĂƌŝĂƟŽŶ ĞƐĐƌŝƉƟŽŶ͗ΗĐŚĂŶŐĞǁŚĞƌĞŝŶƐƉĞĐŝĮĐƐĞƋƵĞŶĐĞ ĐŽŵƉĂƌĞĚƚŽƚŚĞƌĞĨĞƌĞŶĐĞƐĞƋƵĞŶĐĞ͘͘͘Η

/ŶǀĞƌƐŝŽŶ DŽƌĞƚŚĂŶŽŶĞŶƵĐůĞŽƟĚĞƌĞƉůĂĐĞƐƚŚĞŽƌŝŐŝŶĂůƐĞƋƵĞŶĐĞ͕

ďĞŝŶŐŝŶƚŚĞƌĞǀĞƌƐĞͲĐŽŵƉůĞŵĞŶƚ͘

dƌĂŶƐůŽĐĂƟŽŶ ZĂŶŐĞŽĨŶƵĐůĞŽƟĚĞƐĨƌŽŵŚŽŵŽůŽŐŽƵƐĐŚƌŽŵŽƐŽŵĞƌĞƉůĂĐĞ ŽƌŝŐŝŶĂůƐĞƋƵĞŶĐĞ͘

ŽƉLJEƵŵďĞƌ

sĂƌŝĂƟŽŶ sĂƌŝĂďůĞŶƵŵďĞƌŽĨƚĂŶĚĞŵƌĞƉĞĂƚƐĂƌĞŝŶƐĞƌƚĞĚ͘

^ƵďƐƟƚƵƟŽŶ KŶĞŶƵĐůĞŽƟĚĞŝƐƌĞƉůĂĐĞĚďLJĂŶŽƚŚĞƌŶƵĐůĞŽƟĚĞ͘

ĞůĞƟŽŶ KŶĞŽƌŵŽƌĞŶƵĐůĞŽƟĚĞƐĂƌĞĚĞůĞƚĞĚ͘

/ŶƐĞƌƟŽŶ KŶĞŽƌŵŽƌĞŶƵĐůĞŽƟĚĞƐĂƌĞŝŶƐĞƌƚĞĚŝŶƚŚĞƐĞƋƵĞŶĐĞ͕ƚŚĞLJ ĂƌĞŶŽƚĂĐŽƉLJŽĨƚŚĞϱΖ͘

ƵƉůŝĐĂƟŽŶ ĐŽƉLJŽĨŽŶĞŽƌŵŽƌĞŶƵĐůĞŽƟĚĞƐŝƐŝŶƐĞƌƚĞĚϯΖŽĨƚŚĞ ŽƌŝŐŝŶĂůĐŽƉLJ͘

ŽŶǀĞƌƐŝŽŶ ZĂŶŐĞŽĨŶƵĐůĞŽƟĚĞƐĨƌŽŵŽƚŚĞƌůŽĐƵƐƌĞƉůĂĐĞŽƌŝŐŝŶĂů sequence, in this indel.

ĞůĞƟŽŶͲ/ŶƐĞƌƟŽŶ KŶĞŽƌŵŽƌĞŶƵĐůĞŽƟĚĞƐĂƌĞƌĞƉůĂĐĞĚďLJŽŶĞŽƌŵŽƌĞŽƚŚĞƌ ŶƵĐůĞŽƟĚĞƐ͘

2 A variation may have no difference in the amino acid sequence should

the change be present in a noncoding region. Variation in the noncoding area may participate in transcription regulation at even remote loci, however. This is the case with the lactase-persistence polymorphic variant LCT -13910T that is located in intron 13 of the MCM6 gene, more than ͳ͵ԜͲͲͲ „ƒ•‡ ’ƒ‹”• ˆ”‘ –Š‡ ‹‹–‹ƒ–‹‘ •‹–‡ ‘ˆ –Š‡LCT gene that encodes lactase.²⁸ A synonymous variation resides in the exon of a gene but does not alter the amino acid sequence. A nonsynonymous variation alters the amino acid sequence and, hence, changes the protein code.²¹ In addition, a variation may be described according to its potency to cause a loss of function or gain of function. In Table 2 the most common consequences of a given variation are listed.²⁹

Table 2.WŽƐƐŝďůĞƚLJƉĞƐŽĨĐŽŶƐĞƋƵĞŶĐĞƐŽĨǀĂƌŝĂƟŽŶ͘

dLJƉĞŽĨǀĂƌŝĂƟŽŶ ĞƐĐƌŝƉƟŽŶ ĞůĞƟŽŶŽƌĂĚƵƉůŝĐĂƟŽŶ

ŽĨĂŐĞŶĞŽƌĂŶĞdžŽŶ

KŌĞŶƉĂƚŚŽŐĞŶŝĐ͕ĚƵƉůŝĐĂƟŽŶĐŚĂŶŐĞƐĂŵŽƵŶƚŽĨ ƉƌŽƚĞŝŶƉƌŽĚƵĐƚ͖ŵĂƌŬĞĚǀĂƌŝĂƟŽŶďĞƚǁĞĞŶŐĞŶĞƐ͘

sĂƌŝĂƟŽŶŝŶƚŚĞƉƌŽŵŽƚĞƌ

ĂƌĞĂ;ĐŝƐͲĂĐƟŶŐĞīĞĐƚƐͿ WŽƐƐŝďůĞĞīĞĐƚŽŶƚƌĂŶƐĐƌŝƉƟŽŶĂĐƟǀŝƚLJ͘

sĂƌŝĂƟŽŶƚŚĂƚĂīĞĐƚƐ ƐƉůŝĐŝŶŐ

ŚĂŶŐĞƐŝŶƚŚĞĞdžŝƐƟŶŐƐƉůŝĐĞĐŝƚĞƐŽƌĐƌĞĂƟŽŶŽĨ ĐƌLJƉƟĐŽŶĞƐ͖ĐŚĂŶŐĞŝŶƚŚĞƉƌŽƚĞŝŶƉƌŽĚƵĐƚ͘WŽƐƐŝďůĞ ůŽƐƐŽĨĨƵŶĐƟŽŶ͘

&ƌĂŵĞƐŚŝŌǀĂƌŝĂƟŽŶ /ŶƐĞƌƟŽŶŽƌĚĞůĞƟŽŶƚŚĂƚĂůƚĞƌƐƚŚĞƚƌŝƉůĞƚƌĞĂĚŝŶŐ ĨƌĂŵĞ͖ŶŽƉƌŽƚĞŝŶŝƐƉƌŽĚƵĐĞĚ͘>ŽƐƐŽĨĨƵŶĐƟŽŶ͘

EŽŶƐĞŶƐĞǀĂƌŝĂƟŽŶ WƌĞŵĂƚƵƌĞŝŶƚĞƌƌƵƉƟŽŶŽĨƚƌĂŶƐĐƌŝƉƟŽŶ͖ŶŽƉƌŽƚĞŝŶŝƐ ƉƌŽĚƵĐĞĚ͘>ŽƐƐŽĨĨƵŶĐƟŽŶ͘

DŝƐƐĞŶƐĞǀĂƌŝĂƟŽŶ KŶĞĂŵŝŶŽĂĐŝĚŝƐƌĞƉůĂĐĞĚďLJĂŶŽƚŚĞƌ͖ĐŚĂŶŐĞŝŶ ƚŚĞƉƌŽƚĞŝŶƉƌŽĚƵĐƚ͘WŽƐƐŝďůĞůŽƐƐŽƌŐĂŝŶŽĨĨƵŶĐƟŽŶ͘

2.2 IDENTIFYING SUSCEPTIBILITY VARIANTS FOR COMPLEX DISEASES

2.2.1 Heritability

The overall variance of the phenotype (V_P) is the sum of variation due to individual causes:

V_P=V_A+V_D+V_I+V _E

Here V_A is the variance due to additive genetic effects, V_D is the variance due to dominant genetic effects, V_I is the variance of interaction of the effect of genes at several loci, and V_E is the environmental variance.

Heritability (h²) is the extent to which genetic variation (V_G) between individuals explains the difference in the variation of phenotypic trait (V_P).³⁰Š—•ǡ‹–‹•†‡ϐ‹‡†ƒ•ǣ

h²=V_౲/V_P

Traditionally, this narrow-sense heritability was estimated on the basis of closely related individuals, but recently it has become possible to estimate single nucleotide polymorphism (SNP) -based heritability for a complex trait with the variance of unrelated individuals.³¹ Complex diseases arise due to several factors and dichotomizing variance into genetic and environmental has proven over simplistic.³⁰ No knowledge regarding heritability in AKI exists. Regarding complex traits, heritability was deemed missing because the bulk of causative variants were not found with genome wide association analyses.^32,33 Possible explanations for the missing heritability include: variants that are too rare to be detected in common variant genotyping arrays; structural variants; low power to detect gene-gene and gene-environment interactions; low power to detect very small effects of causal common variants; epigenetic effects;

and the underestimation of the heritability because of tagged SNPs used in association studies.^32,33 The effect of shared family environment and assortative mating may lead to overestimation of heritability, suggesting there is no missing heritability in complex diseases.^32,33

Ϯ͘Ϯ͘ϮŽŵƉůĞdžĚŝƐĞĂƐĞƐĂŶĚƚƌĂŝƚƐ

Human appearance and behavior is the result of hundreds of complex

2

&ŝŐƵƌĞ Ϯ͘ WŽƐƐŝďŝůŝƟĞƐ ŝŶ ŝĚĞŶƟĨLJŝŶŐ ŐĞŶĞƟĐ ǀĂƌŝĂŶƚƐ ĂĐĐŽƵŶƟŶŐ ĨŽƌ ƚŚĞ ƌŝƐŬ ĂůůĞůĞ ĨƌĞƋƵĞŶĐLJ;džͲĂdžŝƐͿĂŶĚƐƚƌĞŶŐƚŚŽĨŐĞŶĞƟĐĞīĞĐƚ;LJͲĂdžŝƐͿ͘/ŶƚŚĞůŽǁĞƌƌŝŐŚƚĐŽƌŶĞƌĂƌĞ ƚŚĞ ĐŽŵŵŽŶ ǀĂƌŝĂŶƚƐ ƚŚĂƚ ŵŽƐƚ ŽŌĞŶ ĂƌĞ ŝĚĞŶƟĮĞĚ ǁŝƚŚ ŐĞŶŽŵĞ ǁŝĚĞ ĂƐƐŽĐŝĂƟŽŶ

;'tͿ ƐƚƵĚŝĞƐ ŝŶ ĂƐƐŽĐŝĂƟŽŶ ǁŝƚŚ ĐŽŵŵŽŶ ĚŝƐĞĂƐĞƐ͘ &ŝŐƵƌĞ ĂĚĂƉƚĞĚ ĨƌŽŵ ŽƌŝŐŝŶĂů ƉƵďůŝĐĂƟŽŶƐďLJDĐĂƌƚŚLJ³⁷ĂŶĚDĂŶŽůŝŽĞƚĂů͘³²

thus appearing as missing heritability.³⁵ In Figure 2 the relation between allele frequencies and effect sizes is illustrated.

Most diseases are complex as opposed to Mendelian.²¹ Whereas a Mendelian disease is caused by a single genetic mutation in one or both of the alleles of a causative gene, complex diseases are products of a repertory of susceptibility factors with a small increase to the overall risk. Even if the effect of an individual genetic variant is small, combining the effects of several variants that associate with a trait may be used in a polygenic risk score.¹¹ The genetic variants may be detrimental or protective in relation to the disease. Patterns of inheritance for a variant sequence are regarded identically in both Mendelian and complex disease: in dominant inheritance a single risk allele causes the risk addition, whereas in recessive inheritance both alleles must be risk alleles in order for the risk to confer.

In codominant inheritance, the contribution of the alleles is additive.²¹ Evidence of genetic association in relation to complex diseases has existed for several decades.^8,38 Some established associations are with migraine^9,39,40 and diabetes.^8,41 Figure 3 describes the heritability of some investigated complex traits.

Very rare alleles ĐĂƵƐŝŶŐŵĞŶĚĞůŝĂŶ

diseases

KŶůLJĨĞǁĞdžĂŵƉůĞƐ ŽĨĐŽŵŵŽŶǀĂƌŝĂŶƚƐ ǁŝƚŚŚŝŐŚĞīĞĐƚŽŶ

ĐŽŵŵŽŶĚŝƐĞĂƐĞ

Very rare alleles ǁŝƚŚůŽǁĞīĞĐƚƐŝnjĞ

ƚŚĂƚĂƌĞĚŝĸĐƵůƚ ƚŽŝĚĞŶƟĨLJ

Variants with low frequency ĂŶĚŝŶƚĞƌŵĞĚŝĂƚĞ

ĞīĞĐƚ

ŽŵŵŽŶǀĂƌŝĂŶƚƐ ŝĚĞŶƟĮĞĚǁŝƚŚ't

ŝŶĂƐƐŽĐŝĂƟŽŶǁŝƚŚ ĐŽŵŵŽŶĚŝƐĞĂƐĞ īĞĐƚƐŝnjĞ

ůůĞůĞĨƌĞƋƵĞŶĐLJ ,ŝŐŚ

Low

Rare ŽŵŵŽŶ

&ŝŐƵƌĞϯ͘ƐƟŵĂƚĞƐŽĨŚĞƌŝƚĂďŝůŝƚLJ;Ś²ͿƉĞƌĐĞŶƚĂŐĞĨŽƌĐŽŵƉůĞdžƚƌĂŝƚƐĂƐƉƌĞƐĞŶƚĞĚ ŝŶƉƌĞǀŝŽƵƐƐƚƵĚŝĞƐĂďŽƵƚĂŐĞͲƌĞůĂƚĞĚŵĂĐƵůĂƌĚĞŐĞŶĞƌĂƟŽŶ͕⁴²ƐLJƐƚĞŵŝĐůƵƉƵƐ ĞƌLJƚŚĞŵĂƚŽƐƵƐ͕⁴³ƚLJƉĞϮĚŝĂďĞƚĞƐ͕⁴⁴ĂŶĚ,>ĐŚŽůĞƐƚĞƌŽů͘^ϰϱ

,>͕ŚŝŐŚĚĞŶƐŝƚLJůŝƉŽƉƌŽƚĞŝŶ͘

Ϯ͘Ϯ͘ϯ'ĞŶĞƟĐƐƚƵĚLJĚĞƐŝŐŶƐ

Ϯ͘Ϯ͘ϯ͘ϭ>ŝŶŬĂŐĞƐƚƵĚŝĞƐ

In linkage analyses, the genetic markers are followed on their passage down the generations of a pedigree. Rather large families in which the

–”ƒ‹–‹••‡‰”‡‰ƒ–‡†ƒ”‡”‡“—‹”‡†ˆ‘”Ž‹ƒ‰‡ƒƒŽ›•‹•ǤŠ‡˜‡”›ϐ‹”•–†‹•‡ƒ•‡Ǧ causing genes were located with linkage design. The aim is to create a map of genetic signposts that are in close proximity to the disease-causing defect.²¹Š‡’”‘„ƒ„‹Ž‹–›‘ˆ”‡…‘„‹ƒ–‹‘„‡–™‡‡–™‘Ž‘…‹‹•“—ƒ–‹ϐ‹‡†

by their proximity and a logarithm of the odds (LOD) score is derived. When large, the LOD score indicates that the two loci are linked.²¹ Nonetheless, identifying susceptibility factors for complex diseases by using linkage analysis has shown little success.³³ Nowadays linkage analyses are less frequently used.

Ϯ͘Ϯ͘ϯ͘ϮƐƐŽĐŝĂƟŽŶƐƚƵĚŝĞƐ

In association studies, the hypothesis prevails that the genetic variant of 60

50 40 30 20 10

0 ŐĞͲƌĞůĂƚĞĚŵĂĐƵůĂƌ ĚĞŐĞŶĞƌĂƟŽŶ

^LJƐƚĞŵŝĐůƵƉƵƐ ĞƌLJƚŚĞŵĂƚŽƐƵƐ

dLJƉĞϮ diabetes

HDL cholesterol Heritability (h2, %)

2 in a sample of cases and controls for a given trait.²¹ The hypothesis can

be formulated according to experimental data and theory; however, even in the case of an evident association with the trait, causation cannot be assumed.²¹ It is of vital importance to replicate these kinds of results in an independent sample.²¹ The typical effect size of a risk allele detected to have an association with a trait is 1.1 for heterozygote genotypes and 1.5- 1.6 for homozygote genotypes³⁴. However, with larger sample sizes and the ability to combine the datasets in a meta-analysis, even the variants that confer a very small risk can be detected.

Genome wide association studies (GWAS) screen a plethora of known variants within the genome in association to a trait, and thus provide data about associated variants without a preset hypothesis. The published GWAS studies can be searched by using a GWAS Catalog created by the National Human Genome Research Institute and the European Bioinformatics Institute (https://www.ebi.ac.uk/gwas/).⁴⁶ In April 2019, the catalog …‘–ƒ‹‡†͵Ԝͻʹ͵’—„Ž‹…ƒ–‹‘•™‹–ŠͺͺԜͲͻͷ•ƒ†ͳ͵ͶԜ͹Ͳͷƒ••‘…‹ƒ–‹‘•Ǥ The enabling resource behind hundreds of GWAS studies has been the UK biobank that offers open access to the genomic and health data of more

–ŠƒͷͲͲԜͲͲͲ‹†‹˜‹†—ƒŽ•Ǥ^47,48

Ϯ͘Ϯ͘ϯ͘ϯEŽǀĞůƐĞƋƵĞŶĐŝŶŐƚĞĐŚŶŽůŽŐŝĞƐ

In addition to revolutionizing developments in sequencing techniques, it

Šƒ• „‡…‘‡ ’‘••‹„Ž‡ –‘ ’”‘†—…‡ †ƒ–ƒ ‘”‡ ‡ˆϐ‹…‹‡–Ž› ƒ† ˆ‘” ƒ •ƒŽŽ‡”

cost than previously, hence the scale of sequencing has grown. Novel high-throughput sequencing methods are utilized in, for example, whole- genome, targeted-DNA, whole-transcriptome, and DNA-methylation-site sequencing.⁴⁹ The applications include the comparison of genomes of individuals globally and across human history, as well as targeted according to tissues or individual cells.⁴⁹ The usefulness of exome sequencing for diagnostic purposes has been supported,⁵⁰ moreover in association to kidney disease.⁵¹

Ϯ͘ϯhd</Ez/E:hZz;</Ϳ

Ϯ͘ϯ͘ϭZĞŶĂůĨƵŶĐƟŽŶ

The kidneys are organs with multiple functions of vital importance. In order to achieve body homeostasis, the kidneys must remove uremic retention solutes and foreign substances, regulate the balance of electrolytes and water in respect to plasma osmolality, and maintain acid-base balance.⁵² In addition, kidneys participate in regulating systemic blood pressure, stimulation of red blood cell production, gluconeogenesis, and synthesis of the active form of vitamin D.⁵²

‹†‡›•”‡…‡‹˜‡ʹͲΨ‘ˆ–Š‡”‡•–‹‰…ƒ”†‹ƒ…‘—–’—–ˆ‘”ϐ‹Ž–”ƒ–‹‘Ǥ⁵³ For a healthy individual with two kidneys, the functional capacity of the kidneys is adequate in most situations. Each kidney contains approximately 1 million functional units — nephrons.⁵² The ability of the kidneys to remove

™ƒ•–‡’”‘†—…–•‹•‘–…‘’”‘‹•‡†—–‹Žƒ•‹‰‹ϐ‹…ƒ–Ž‘••‘ˆ‡’Š”‘•†—‡

to injury.^54,55

Ž‘‡”—Žƒ”ϐ‹Ž–”ƒ–‹‘”ƒ–‡ȋ Ȍ†‡•…”‹„‡•–Š‡ϐ‹Ž–”ƒ–‡˜‘Ž—‡…”‡ƒ–‡†‹

a nephron over time.⁵² It is usually expressed as the rate in proportion to standardized body surface area.⁵⁶ Measuring GFR in an individual requires techniques that are invasive and sometimes cumbersome.^56,57 Creatinine is produced in the body from muscle creatine phosphate.⁵⁸‹†‡›•ϐ‹Ž–‡”

and secrete creatinine. In clinical use, serum creatinine is utilized for a computational estimate of GFR (eGFR). Such estimation according to an endogenous marker is feasible in a steady state^54,59 and is affected by factors besides GFR, such as tubular secretion and extrarenal elimination.⁵⁹ Formulas developed for this purpose are created and validated.^59–64 In addition to using serum creatinine as a surrogate marker for renal function, they account for variables such as age, gender, and race.^62,63 In AKI, neither the production nor elimination of creatinine are in steady state, and, thus, computational creatinine based eGFR formulas are not feasible.

Urine output has an indirect relation to GFR and thus cannot be regarded as an indicative measure of kidney function.^1,65 However, oliguria is frequent among the critically ill.^65,66 Using merely plasma creatinine for AKI diagnosis decreases the sensitivity for detecting AKI⁶⁵ and underestimates the severity of outcomes.⁶⁶

Ϯ͘ϯ͘Ϯ</ĚĞĮŶŝƟŽŶ

AKI syndrome encompasses the multiple etiologies that cause renal function to deteriorate over short period of time.¹ Patients with AKI present with accumulation of waste products, such as creatinine, and reduction in urine output.⁶⁷ Whether this is a sign of injury of the kidney tissue or impairment of the function in relation to the demand at hand, AKI is associated with decreased survival in mild to severe stages.^68,69

Š‡Žƒ–‡•–…‘•‡•—•†‡ϐ‹‹–‹‘ˆ‘”™ƒ•’—„Ž‹•Š‡†‹ʹͲͳʹ„›–Š‡

Acute Kidney Injury Work Group,¹ and was based on the previously validated Risk, Injury, Failure, Loss, End-stage disease (RIFLE) criteria⁷⁰ by the Acute Dialysis Quality Initiative (ADQI), and Acute Kidney Injury Network (AKIN) …Žƒ••‹ϐ‹…ƒ–‹‘Ǥ⁷¹ ‡ˆ‘”‡ –Š‡ •›•–‡ ˆ‘” †‹ƒ‰‘•‹• ƒ† …Žƒ••‹ϐ‹…ƒ–‹‘

2

‹…‡–Š‡™‘”‘ˆ–Š‡ǡ–Š‡…Žƒ••‹ϐ‹…ƒ–‹‘™ƒ•’—„Ž‹•Š‡†„›–Š‡

Acute Kidney Injury Network in an attempt to better discover the very

•ƒŽŽ…Šƒ‰‡•‹…”‡ƒ–‹‹‡ǡƒ†–‘‰‡‡”ƒ–‡ƒ‘”‹‰‹ƒŽ†‡ϐ‹‹–‹‘ˆ‘”–Š‡

syndrome with a spectrum wider than mere renal failure.⁷¹ However, both of these existing criteria were proven inadequate as they disregarded relevant patients.⁷⁵ By creating the clinical practice guideline for evaluation and management of AKI, the Acute Kidney Injury Work Group generated

ƒ —‹ˆ‘” †‡ϐ‹‹–‹‘ ƒ† ‰”ƒ†‹‰ •›•–‡ –‘ •—„•–‹–—–‡ –Š‡ ’”‡˜‹‘—•

†‡ϐ‹‹–‹‘•Ǥ¹

‹•…Žƒ••‹ϐ‹‡†‹–‘•‡˜‡”‹–›•–ƒ‰‡•ƒ……‘”†‹‰–‘…”‹–‡”‹ƒǤ¹ This …Žƒ••‹ϐ‹…ƒ–‹‘”ƒ‰‡•ˆ”‘•–ƒ‰‡ͳ–‘͵Ǥ‘”‡•‡˜‡”‡‹•ƒ••‘…‹ƒ–‡†™‹–Šƒ

worse outcome.^14,15,76 However, clinical judgement and interest towards the cause of AKI are warranted in addition to preset thresholds of diagnosis and stage of severity.¹

Ϯ͘ϯ͘ϯWĂƚŚŽƉŚLJƐŝŽůŽŐLJŽĨ</

As AKI is a complex syndrome presenting in differing clinical scenarios, no uniform pathophysiology exists.^6,7 Investigation of the pathology in

ƒ‹ƒŽ‘†‡Ž•Šƒ•—†‡”Ž‹‡†–Š‡†‹ˆϐ‹…—Ž–‹‡•‹‹‹…‹‰–Š‡ƒŽ–‡”ƒ–‹‘

in physiology in a different species.^55,77,78 However, common underlying injury processes of cellular depolarization, apoptosis, and necrosis,⁵⁵ as

™‡ŽŽ ƒ• ‹ϐŽƒƒ–‹‘⁷⁸ have been found to follow all forms of insult. In depolarization, the cells of the tubular endothelium lose their ability to function, but recovery is possible.⁵⁵ When more severe damage is encountered, the cells die either by energy consuming apoptosis or necrosis.⁵⁵

Sepsis is the most common underlying condition of acute deterioration of renal function in the ICU setting.^15,79,80 In a Finnish cohort of critically ill patients, the AKI incidence in septic patients was 53%.¹⁶ Sepsis-associated AKI (SA-AKI) is unique from AKI without sepsis in pathophysiology, timing, and outcomes.⁸⁰ However, our understanding of the SA-AKI pathophysiology is perpetually evolving.^7,80,81

The assumption of global renal hypoperfusion preceding septic AKI has been invalidated.^82,83 Moreover, renal circulation appears to be uncoupled form systemic hemodynamics and, thus, cannot be reliably predicted from it.⁸⁴ In addition, oxygen-dependent metabolism is globally preserved in animals with septic shock and AKI.⁸⁵ ‘’Ž‡š ‹–”ƒǦ”‡ƒŽ „Ž‘‘† ϐŽ‘™

changes suggests that regional distribution in microcirculation, along with patency and caliber of renal vessels, plays a role in AKI development.^86–88 Renal vascular resistance and interstitial edema due to vascular permeability, as well as vasoconstriction, vasodilation, and oxidative stress -induced endothelial dysfunction are regarded as possible contributors to the development of SA-AKI.⁸⁷

As sepsis entails a dysregulated host response,⁸⁹ the emergence of

ƒŽ–‡”ƒ–‹‘•‹‹ϐŽƒƒ–‹‘’ƒ–Š™ƒ›•ƒ†‹—‘Ž‘‰‹…’”‘ϐ‹Ž‡•ƒŽ‘‰–Š‡

spectrum are plausible.^87,90 In sepsis, the pathogen-associated molecules, and occasionally endogenous molecules from injured cells, are recognized

„› ’ƒ––‡”Ǧ”‡…‘‰‹–‹‘ ”‡…‡’–‘”•ǡ ƒ† ‹ϐŽƒƒ–‘”› ‰‡‡ –”ƒ•…”‹’–‹‘

is up-regulated.^87,91 A storm of cytokines follows, during which pro-

‹ϐŽƒƒ–‘”› ‡†‹ƒ–‘”• ƒ…–‹˜ƒ–‡ ‡†‘–Š‡Ž‹ƒŽ …‡ŽŽ• ƒ† ‹…”‡ƒ•‡ ˜ƒ•…—Žƒ”

permeability.⁸⁷ More leucocytes are recruited to the kidney by adhesion molecules on the endothelium.^87,88 The function of neutrophilic leucocytes

ˆ—”–Š‡” ’”‘‘–‡• –Š‡ ‹ϐŽƒƒ–‹‘ ”‡ƒ…–‹‘ǡ ™Š‹…Š ‹• ”‡•’‘•‹„Ž‡ ˆ‘”

some tissue damage.⁸⁷Š‡‹ϐŽƒƒ–‘”›‹Œ—”‹‡•‹‡†‘–Š‡Ž‹ƒŽƒ†”‡ƒŽ

tubular cells are considered key in the initiation of AKI.⁹²

Apoptosis is one suggested consequence of cellular injury in SA-AKI.^80,93 However, the histologic changes in SA-AKI have been heterogeneously distributed and without marked signs of apoptosis or necrosis.77,87,88,94,95 In addition, vacuolization is evident in these tubular cells that are subjected to oxidative stress, which is infrequent in SA-AKI.⁸⁷ They adapt to the lack of ATP (adenosine triphosphate) by entering cell-cycle arrest with orchestration from the mitochondria.^87,88 Subsequently, energy is withheld from futile functions, rendering renal function reduced.^87,88 Once injured, the tubular endothelium is dysfunctional for sodium reabsorption, thus triggering tubuloglomerular feedback, which by decreasing hydrostatic pressure in the glomerulus decreases GFR.⁸⁸

Major surgery is another substantial cause of AKI in hospitalized patients with varying incidence according to type of procedure.^7,96–98 The underlying etiological factors vary accordingly, including alterations in hemodynamics, exposure to toxins, ischemia-reperfusion injury, embolization of the renal artery, neurohormonal responses to hypotension

ƒ†–‹••—‡‹Œ—”›ǡ‹ϐŽƒƒ–‹‘ǡƒ†‘š‹†ƒ–‹˜‡•–”‡••Ǥ⁷ A designated entity is cardiac surgery-associated AKI (CSA-AKI), which is the second most common type of AKI in the ICU setting.⁹⁹ The pathophysiology of CSA-AKI is multifactorial and incompletely understood.⁹⁹ When cardiopulmonary bypass (CPB) is used, the perfusion is non-pulsatile, there is hemodilution and temperature changes.⁹⁹ Intravascular hemolysis can develop because of CPB and cause tubular epithelial cell injury when free hemoglobin is present.⁹⁹ After CPB, there is a risk of ischemia-reperfusion injury, with

’”‘†—…–‹‘‘ˆ”‡ƒ…–‹˜‡‘š›‰‡•’‡…‹‡•ƒ†‹ϐŽƒƒ–‹‘Ǥ⁹⁹

Nephrotoxic exposure is the predisposing factor in 15% to 19% of ICU patients with AKI.^15,79 Biological effects vary between nephrotoxins, which can be both exogenous and endogenous.⁷ In cisplatin-induced injury the

2 media.¹⁰¹ The pathophysiology of CI-AKI is recognized to include both toxic

effects on tubular cells and hemodynamic alterations, resulting in reduced renal perfusion.¹⁰¹ Excess accumulation of endogenous substances, such as myoglobin which is released to bloodstream in rhabdomyolysis, contribute to AKI formation in mechanisms that may include oxidative stress and vascular alterations.^102,103

The necessity of classifying AKI into clinical sub-phenotypes has been claimed.⁷•—„Ǧ’Š‡‘–›’‡•™‹–Š†‹•–‹…–‘—–…‘‡•Šƒ˜‡„‡‡‹†‡–‹ϐ‹‡†

using a latent class-analysis approach that modelled predetermined baseline clinical data and circulating plasma biomarkers involved in

‡†‘–Š‡Ž‹ƒŽ†›•ˆ—…–‹‘ǡƒ†‹ϐŽƒƒ–‹‘ƒ†ƒ’‘’–‘•‹•’ƒ–Š™ƒ›•Ǥ¹⁰⁴ Of note, the potentially reversible causes of AKI, such as renal artery

‘……Ž—•‹‘ǡ ‹ϐŽƒƒ–‘”› ‹†‡› ’ƒ”‡…Š›ƒŽ †‹•‡ƒ•‡•ǡ ‘” ‘„•–”—…–‹‘ ‹

the urinary tract, should be determined and treated to avoid irreversible damage to the kidneys.¹

2.3.4 Risk factors for AKI

Multiple predisposing factors for AKI are recognized, and these frequently occur coincidentally. The most common underlying insults resulting in ‹ Š‘•’‹–ƒŽǦƒ†‹––‡† ’ƒ–‹‡–• ƒ”‡ •‡’•‹•ǡ ‹ϐŽ—‡…‡ ‘ˆ ƒ ‡’Š”‘–‘š‹…

agent, and reduced circulation to the kidneys in situations such as major surgery.¹⁵ In addition, several susceptibilities can increase the risk of AKI.

Chronic comorbidities are associated with increased risk of developing AKI. The most evidence to this extent is in relation to chronic kidney disease (CKD), which is an independent risk factor for AKI.^1,14,105 Patients with diabetes have an increased risk of AKI.^1,106,107 AKI is more common in patients with liver failure¹⁰⁸ or heart failure.¹⁰⁷ Patients with cancer are at increased risk for AKI, partially because of tumor lysis syndrome in association with a high tumor burden or cell turnover.¹⁰⁹

Advanced age is associated with higher risk of AKI.15,106,107,110 In addition, AKI is more prevalent in people of African descent.^106,111 Pre-ICU hypovolemia has been presented to independently increase risk for AKI.^1,14

‘™‡˜‡”ǡ…‘‰‡•–‹‘†—‡–‘‡š…‡••‹˜‡ϐŽ—‹†ƒ†‹‹•–”ƒ–‹‘Šƒ•ƒŽ•‘„‡‡

highlighted.¹¹² High risk surgery and emergency surgery increase the risk of AKI.¹⁰⁷ Nephrotoxicity by drugs such as antibiotics and diuretics,14,106,107,110

and pre-ICU use of colloids¹⁴ have been found to predispose to AKI. Female

‰‡†‡”Šƒ•„‡‡‹†‡–‹ϐ‹‡†ƒ•ƒ”‹•ˆƒ…–‘”ˆ‘”ǡ¹ however recently this

Šƒ•„‡‡†‡„ƒ–‡†‹”‡Žƒ–‹‘–‘…‘ϐŽ‹…–‹‰”‡•—Ž–•‹‡š’‡”‹‡–ƒŽƒ‹ƒŽ

models.¹¹³

Early recognition of AKI risk is advantageous in possibly preventing the onset of the condition. Thus, risk prediction tools that combine clinical data have been researched.¹¹⁴ In developing a risk prediction model, the challenges lie in choosing the predictors and evaluating how the model

performs.¹⁰ The models developed thus far^115–117 have provided average performance (positive predictive value range from 23% to 38%) and the chosen predictors are heterogeneous.¹¹⁴

Ϯ͘ϯ͘ϱ</ĞƉŝĚĞŵŝŽůŽŐLJ

AKI syndrome is a frequent complication to other conditions, especially in critically ill patients.¹¹⁸ In critically ill patients the AKI incidence is 36- 67%.14,15,69,73,106 The global incidence of AKI has been steadily increasing.¹¹⁹ In hospitalized adult patients, the pooled incidence is approximately 20%.^12,13 In low-to-middle-income countries, the patients are generally younger and AKI is the result of a single noxa, whereas in high-income countries the patients are older and commonly have an underlying severe illness.^118,120 In addition, the treatment modalities and delay to their onset frequently hampers the care of AKI in low-to-middle-income countries.^118,121

AKI is associated with adverse outcomes, both short and long term, in regard to patients and the kidneys.¹¹⁸ AKI is an independent predictor of in-hospital mortality,15,69,73,108 as well as mortality at 90 days^14,122,123 and at 1 year.14,106,122,124 The long term survival after AKI has been deemed poor,¹²⁵ as the patients are at increased risk for death after discharge from ICU and hospital.¹⁰⁵ This risk is, however, mostly determined by age and pre- existing comorbidities,¹²⁶ and even a dialysis-requiring AKI episode may not independently hamper long-term survival.¹²⁷ For AKI survivors, the perceived health-related quality-of-life (HRQL) is acceptable or similar in comparison to survivors without AKI,^105,128 and renal recovery happens in the majority of the cases.¹⁰⁵ Although the HRQL decreases in relation to the general population, most AKI survivors would undergo similar treatment again if needed.¹²⁸

In order to recover from AKI, kidney tubular epithelium is mandated to undergo proliferation, in which stem cells and growth factors are denoted in animal models.⁵⁵ The degree of renal recovery can be assessed by determining renal functional reserve, which describes the capacity of

–Š‡‹†‡›•–‘‹…”‡ƒ•‡ ƒ•‡‡†‡†„›Š›’‡”ϐ‹Ž–”ƒ–‹‘‹–Š‡”‡ƒ–

nephrons.⁵⁴ When renal recovery is not achieved, patients remain dialysis dependent. In addition, distinct recovery phenotypes are associated with long term outcomes.⁵ Risk of requiring maintenance dialysis is increased when renal replacement therapy (RRT) is necessitated in the acute phase.¹²⁷ Moreover, AKI is associated with increased risk of chronic kidney disease and end-stage renal failure.^122,129

2 2 Ϯ͘ϯ͘ϲ</ƚƌĞĂƚŵĞŶƚŵŽĚĂůŝƟĞƐ

No known curative treatment for AKI exists.¹³³ Hence, prevention is of vital importance. Known risk factors confer a major proportion, especially in the younger age groups of the critically ill.¹⁰⁶ Avoidance of nephrotoxic agents

„›‹†‹˜‹†—ƒŽǦŽ‡˜‡Ž•–”ƒ–‹ϐ‹…ƒ–‹‘ƒ†—–‹Ž‹œƒ–‹‘‘ˆ’”‡˜‡–‹˜‡•–”ƒ–‡‰‹‡•‹•

suggested.¹⁰¹

Optimizing hemodynamics and volume status is a well-intentioned

‰‘ƒŽǡ Š‘™‡˜‡” Šƒ”† –‘ ƒ……‘’Ž‹•Š ™‹–Š –Š‡ †‹ˆϐ‹…—Ž– –ƒ• ‘ˆ †‡–‡”‹‹‰

optimal for existing kidney circumstance. Fluid overload is potentially harmful for critically ill patients, as is untreated hypovolemia.^134,135 The optimal solution to replenish the volume shortage has been under debate, with the strongest evidence for the majority of patients in favor of buffered salt solutions, and against synthetic colloids, that have been proven unsafe and not superior to crystalloids.^134,135

For established AKI, the pharmacological interventions have generally proven otiose. Diuretics are not recommended in prevention or treatment of AKI.^1,133 Alkaline phosphatase has been experimentally administered

–‘ •‡’–‹… ’ƒ–‹‡–• ™‹–Š …‘ϐŽ‹…–‹‰ ‡ˆˆ‡…–• ‘ ”‡ƒŽ ˆ—…–‹‘Ǥ^87,136,137 No medication is accepted for clinical use in AKI treatment. Elimination of

‡†‘–‘š‹ƒ†‹ϐŽƒƒ–‘”›‡†‹ƒ–‘”•‹Ǧ„›Š‡‘Ǧƒ†•‘”’–‹‘Šƒ•

been experimentally established.⁸⁷

The use of renal replacement therapy (RRT) in the critically ill is common.¹⁵ Several uncertainties remain, nonetheless, about the optimal administration and timing of RRT in AKI.¹³⁸ The conventional indications of initiating RRT include acidosis, hyperkalemia, uremic symptoms, oliguria or anuria, and volume overload; however, it is suggested that initiation of

–Š‡–”‡ƒ–‡–‡˜‡„‡ˆ‘”‡ƒ’’‡ƒ”ƒ…‡‘ˆ–Š‡•‡‹‰Š–„‡„‡‡ϐ‹…‹ƒŽǤ¹³⁹

3 AIMS OF THE STUDY

The objective of this study was to investigate whether selected genetic variants associate with the risk to develop acute kidney injury in adult patients in an intensive care unit setting.

Š‡•’‡…‹ϐ‹…ƒ‹•‘ˆ–Š‹••–—†›™‡”‡ǣ

1. To systematically review the current literature for genetic predisposition to AKI (I)

2. To evaluate the quality of published studies (I)

3. ‘…‘ϐ‹”‘””‡ˆ—–‡–Š‡’”‡˜‹‘—•ϐ‹†‹‰•”‡‰ƒ”†‹‰•‡Ž‡…–‡†

variants in genes related to 3.1. apoptosis (II)

3.2. iron metabolism (III) 3.3. ‹ϐŽƒƒ–‹‘ȋȌǤ

3, 4

4 MATERIALS AND METHODS

4.1 PATIENTS

Patients in Studies II–IV were from the prospective, observational Finnish Acute Kidney Injury (FINNAKI) study. Seventeen Finnish ICUs took part in this multicenter study between September 1^st, 2011 and February 1^st, 2012. In addition to the main cohort, an extended cohort was enrolled until April 30^th to achieve the desired number of patients with sepsis. For the genetic study, 2968 patients gave their consent. After 122 DNA samples failed in isolation, 2846 patients were included for genetic association analyses. In addition, there were some samples with low success rate in each assay, which led to exclusion of these patients from the analysis.

Patients with emergency ICU admission of any duration or elective admission with an expected stay of more than 24 hours were included in the FINNAKI study. Conversely, patients were excluded from the study if they were under 18 years-of-age, re-admitted with RRT in previous admission, on maintenance dialysis, organ donors, without permanent

”‡•‹†‡…›‹ ‹Žƒ†‘”™‹–Š‹•—ˆϐ‹…‹‡–Žƒ‰—ƒ‰‡•‹ŽŽ•ǡ–”ƒ•ˆ‡””‡†ˆ”‘

another ICU after full (5 days) participation, or in intermediate care. For patients with multiple admissions during the study period, the admission with the highest KDIGO stage was chosen for the study.

For Studies II and III, the patients with known or suspected CKD and KDIGO Stage 1 AKI were excluded; for Study III, only patients with sepsis were included; and for Study IV, the entire cohort was included.

‹‰—”‡ Ͷ ‹ŽŽ—•–”ƒ–‡• ƒ ϐŽ‘™ …Šƒ”– ˆ‘” –—†‹‡• ȂǤ ƒ„Ž‡ ͵ ’”‡•‡–•

patient characteristics for the entire FINNAKI genetic study as well as for Studies II–IV.

&ŝŐƵƌĞϰ͘&ůŽǁĐŚĂƌƚŝůůƵƐƚƌĂƟŶŐƉĂƟĞŶƚŇŽǁŝŶ^ƚƵĚŝĞƐ//ʹ/s͘&/EE</͕&ŝŶŶŝƐŚĐƵƚĞ

<ŝĚŶĞLJ/ŶũƵƌLJ͖E͕ĚĞŽdžLJƌŝďŽŶƵĐůĞŝĐĂĐŝĚ͖<͕ĐŚƌŽŶŝĐŬŝĚŶĞLJĚŝƐĞĂƐĞ͖</'K͕

<ŝĚŶĞLJŝƐĞĂƐĞ͗/ŵƉƌŽǀŝŶŐ'ůŽďĂůKƵƚĐŽŵĞƐ͖</'Kϭ͕<ŝĚŶĞLJŝƐĞĂƐĞ͗/ŵƉƌŽǀŝŶŐ 'ůŽďĂůKƵƚĐŽŵĞƐʹ</^ƚĂŐĞϭ͖</͕ĂĐƵƚĞŬŝĚŶĞLJŝŶũƵƌLJ͘

Ϯϵϲϴ&/EE</ŐĞŶĞƟĐƐƚƵĚLJƉĂƟĞŶƚƐ

ϮϴϰϲƉĂƟĞŶƚƐ

Study IV Study III

1917 no

ƐĞƉƐŝƐ 3 low

success rate

ϲϮϱƉĂƟĞŶƚƐ with AKI ϭϱϳϵƉĂƟĞŶƚƐ

without AKI ϵϮϵƉĂƟĞŶƚƐ

ǁŝƚŚƐĞƉƐŝƐ

86 CKD, known ŽƌƐƵƐƉĞĐƚĞĚ

ϴϰϯƉĂƟĞŶƚƐ ǁŝƚŚƐĞƉƐŝƐ 6 low

success rate

ϯϬϬƉĂƟĞŶƚƐ with AKI ϯϱϯƉĂƟĞŶƚƐ

without AKI 184 KDIGO 1

ϴϯϳƉĂƟĞŶƚƐ ǁŝƚŚƐĞƉƐŝƐ

ϲϱϯƉĂƟĞŶƚƐ ǁŝƚŚƐĞƉƐŝƐ Study II

49 low success rate

ϮϳϵϳƉĂƟĞŶƚƐ

230 CKD, known ŽƌƐƵƐƉĞĐƚĞĚ

ϮϱϲϳƉĂƟĞŶƚƐ

ϮϴϰϯƉĂƟĞŶƚƐ

199 known CKD 440 KDIGO 1

ϮϮϬϰƉĂƟĞŶƚƐ 421 KDIGO 1

ϮϭϰϲƉĂƟĞŶƚƐ

ϲϬϭƉĂƟĞŶƚƐ with AKI ϭϱϰϱƉĂƟĞŶƚƐ

without AKI

ϯϱϰƉĂƟĞŶƚƐ ǁŝƚŚƐĞƉƐŝƐ

ϮϮϲƉĂƟĞŶƚƐŝŶ shock

ϮϱϮƉĂƟĞŶƚƐŝŶ shock ϮϵϵƉĂƟĞŶƚƐ

ǁŝƚŚƐĞƉƐŝƐ

ϭϮϮ;ϰ͘ϭйŽĨϮϵϲϴͿƐĂŵƉůĞƐ ĨĂŝůĞĚŝŶEŝƐŽůĂƟŽŶ

4

ϰ͘ϭ͘ϭŝĂŐŶŽƐƟĐĚĞĮŶŝƟŽŶƐ

ϰ͘ϭ͘ϭ͘ϭĐƵƚĞŬŝĚŶĞLJŝŶũƵƌLJ

…—–‡ ‹†‡› ‹Œ—”› ™ƒ• †‡ϐ‹‡† ƒ• ƒ ƒ„”—’– †‡…Ž‹‡ ‹ ‰Ž‘‡”—Žƒ”

ϐ‹Ž–”ƒ–‹‘ǡƒ•‡˜‹†‡–„›ƒ……——Žƒ–‹‘‘ˆ™ƒ•–‡’”‘†—…–•ƒ††‡…”‡ƒ•‡‹

urine output, according to KDIGO criteria.¹ Plasma creatinine was measured

†ƒ‹Ž›ƒ†—”‹‡‘—–’—–Š‘—”Ž›ǡˆ‘”–Š‡†—”ƒ–‹‘‘ˆϐ‹˜‡†ƒ›•Ǥƒ•‡Ž‹‡

•‡”— …”‡ƒ–‹‹‡ ™ƒ• †‡ϐ‹‡† ƒ• ‘•– ”‡…‡– ‡ƒ•—”‡‡– ™‹–Š‹ –Š‡

previous year, ruling out measurements from the previous week. Missing data for baseline plasma creatinine was addressed as recommended by the ADQI:⁷¹ Assuming normal GFR (75ml/min/1.73 m²), plasma creatinine

™ƒ• …ƒŽ…—Žƒ–‡† ˆ”‘ –Š‡ ‘†‹ϐ‹…ƒ–‹‘ ‘ˆ ‹‡– ‹ ‡ƒŽ ‹•‡ƒ•‡ ȋȌ equation.^60,62 AKI was staged according to disease severity, using the highest resulting stage for each patient. The measurement thresholds for

†‡ϐ‹‹–‹‘•ƒ†•–ƒ‰‡•ƒ”‡’”‡•‡–‡†‹ƒ„Ž‡ͶǤ

Table 3.WĂƟĞŶƚĐŚĂƌĂĐƚĞƌŝƐƟĐƐŝŶƚŚĞĞŶƟƌĞ&/EE</ŐĞŶĞƟĐƐƚƵĚLJ͕ĂƐǁĞůůĂƐ ƉƌĞƐĞŶƚĞĚƐĞƉĂƌĂƚĞůLJĨŽƌ^ƚƵĚŝĞƐ//͕///͕ĂŶĚ/s͘DĞĚŝĂŶ;ŝŶƚĞƌƋƵĂƌƟůĞƌĂŶŐĞͿŝƐŐŝǀĞŶ ĨŽƌĐŽŶƟŶƵŽƵƐǀĂƌŝĂďůĞƐĂŶĚƚŽƚĂůŶƵŵďĞƌ;ƉĞƌĐĞŶƚŽĨĂīĞĐƚĞĚŝŶĂŐƌŽƵƉͿĨŽƌ ĐĂƚĞŐŽƌŝĐĂůǀĂƌŝĂďůĞƐ͘&/EE</͕&ŝŶŶŝƐŚĐƵƚĞ<ŝĚŶĞLJ/ŶũƵƌLJ͖^W^͕^ŝŵƉůŝĮĞĚĐƵƚĞ WŚLJƐŝŽůŽŐLJ^ĐŽƌĞ͖^ƵƌŐŝĐĂůĂĚŵŝƐƐŝŽŶ͕ƐƵƌŐŝĐĂůĂĚŵŝƐƐŝŽŶĚŝĂŐŶŽƐŝƐƌĞŐĂƌĚůĞƐƐŽĨ ƵƌŐĞŶĐLJŽĨĂĚŵŝƐƐŝŽŶ͘

FINNAKI ŐĞŶĞƟĐƐƚƵĚLJ n=2968

Study II

n=2146 Study III

n=653 Study IV

n=2843

ŐĞ͕LJĞĂƌƐ 64 (51–74) 62 (50–72) 63 (53–74) 64 (51–74) 'ĞŶĚĞƌ͕ŵĂůĞ 1894 (63.8) 1348 (62.8) 418 (64.0) 1817 (63.9) ĂƐĞůŝŶĞƉůĂƐŵĂ

ĐƌĞĂƟŶŝŶĞ;ђŵŽůͬůͿ ^82.9 (69.0–96.0) 80.0

(68.0–94.0) 79.0

(67.0–93.3) 82.9 (69.0–95.7) ŵĞƌŐĞŶĐLJ

ĂĚŵŝƐƐŝŽŶ 2603 (88.7) 1907 (89.7) 635 (98.0) 2516 (89.4)

^ƵƌŐŝĐĂůĂĚŵŝƐƐŝŽŶ 1030 (34.7) 709 (33.1) 154 (23.6) 968 (34.1)

SAPS II 37 (28–49) 36 (27–49) 43 (34–55) 37 (28–49)

ϰ͘ϭ͘ϭ͘ϮŚƌŽŶŝĐŬŝĚŶĞLJĚŝƐĞĂƐĞ;<Ϳ

‹•†‡ϐ‹‡†ƒ•ƒ„‘”ƒŽ‹–‹‡•™‹–Š‹‹†‡›•–”—…–—”‡‘”ˆ—…–‹‘ǡ–Šƒ–

are present for a duration of more than 3 months.¹⁴⁰ CKD is diagnosed with either reduced GFR (<60ml/min/1.73 m²) or presence of at least one of the other signs of kidney damage, such as albuminuria, urine sediment abnormalities, electrolyte and other abnormalities due to tubular disorders, abnormal kidney biopsy result, structural abnormalities in an imaging study, or history of kidney transplantation, for this duration.¹⁴⁰ In Studies II to IV, CKD diagnosis was retrieved from the patient medical records, and no retrospective calculations of eGFR were performed. Suspected CKD was considered when a patient received RRT before admission without previous CKD diagnosis.

ϰ͘ϭ͘ϭ͘ϯ^ĞƉƐŝƐ

‡’•‹•™ƒ•†‡ϐ‹‡†ƒ……‘”†‹‰–‘–Š‡ͳͻͻʹ’—„Ž‹•Š‡†…‘•‡•—•…‘ˆ‡”‡…‡ declaration by the American College of Chest Physicians/Society of Critical Care Medicine.¹⁴¹ ……‘”†‹‰ –‘ –Š‡‹” †‡ϐ‹‹–‹‘ǡ •‡’•‹• ‹• –Š‡ •›•–‡‹…

response to infection with two or more manifestations described as

•›•–‡‹…‹ϐŽƒƒ–‘”›”‡•’‘•‡•›†”‘‡ȋȌǤŠ‡•‡ƒ‹ˆ‡•–ƒ–‹‘•

include body temperature >38°C or <36°C, heart rate of >90 beats-per- minute, respiratory rate of >20 breaths-per-minute or partial pressure of

‘š›‰‡‹ƒ”–‡”‹ƒŽ„Ž‘‘†δ͵ʹ‰ǡƒ†™Š‹–‡„Ž‘‘†…‡ŽŽ…‘—–εͳʹԜͲͲͲȀ mm³ or <4000/mm³ or >10% of immature neutrophils. In addition, severe

Table 4͘<ŝĚŶĞLJŝƐĞĂƐĞ͗/ŵƉƌŽǀŝŶŐ'ůŽďĂůKƵƚĐŽŵĞƐ;</'KͿĐƌŝƚĞƌŝĂĨŽƌĚŝĂŐŶŽƐŝŶŐ ĂŶĚƐƚĂŐŝŶŐĂĐƵƚĞŬŝĚŶĞLJŝŶũƵƌLJ;</Ϳ͘¹ŝƚŚĞƌĐƌŝƚĞƌŝĂ;ƐĞƌƵŵͬƉůĂƐŵĂĐƌĞĂƟŶŝŶĞ ŽƌƵƌŝŶĞŽƵƚƉƵƚͿƐƵĸĐĞƐĨŽƌ</ĚŝĂŐŶŽƐŝƐ͘dŚĞĐŚĂŶŐĞŝŶďĂƐĞůŝŶĞƐĞƌƵŵͬƉůĂƐŵĂ ĐƌĞĂƟŶŝŶĞŝƐƉƌĞƐƵŵĞĚƚŽŚĂǀĞŽĐĐƵƌƌĞĚǁŝƚŚŝŶϳĚĂLJƐ͘

^ƚĂŐĞ ^ĞƌƵŵͬƉůĂƐŵĂĐƌĞĂƟŶŝŶĞ hƌŝŶĞŽƵƚƉƵƚ 1 ϭ͘ϱʹϭ͘ϵƟŵĞƐďĂƐĞůŝŶĞ

ORшϮϲ͘ϱђŵŽůͬůŝŶĐƌĞĂƐĞǁŝƚŚŝŶϰϴŚŽƵƌƐ фϬ͘ϱŵůͬŬŐͬŚĨŽƌϲʹϭϮŚŽƵƌƐ 2 Ϯ͘ϬʹϮ͘ϵƟŵĞƐďĂƐĞůŝŶĞ фϬ͘ϱŵůͬŬŐͬŚĨŽƌшϭϮŚŽƵƌƐ

3

ϯ͘ϬƟŵĞƐďĂƐĞůŝŶĞ ORŝŶĐƌĞĂƐĞƚŽшϯϱϯ͘ϲђŵŽůͬů ORŝŶŝƟĂƟŽŶŽĨƌĞŶĂůƌĞƉůĂĐĞŵĞŶƚ ƚŚĞƌĂƉLJ

фϬ͘ϯŵůͬŬŐͬŚĨŽƌшϮϰŚŽƵƌƐ ORĂŶƵƌŝĂĨŽƌшϭϮŚŽƵƌƐ