{"paper_id":"04e81927-4a4f-4316-99ce-6947cd459337","body_text":"1 Anticoagulation-related complications and their outcomes in \n2 hemodialysis patients with acute kidney injury at selected \n3 hospitals in Ethiopia. \n4 Hanan Muzeyin Kedir (hanan.muzeyin@aau.edu.et) 1*, Abdella Birhan Yabeyu  \n5 (birhanabdella@gmail.com)2, Addisu Melkie Ejigu (addisumelkieejigu@gmail.com)3, Tamrat \n6 Assefa Tadesse (tamrat.assefa@aau.edu.et) 1, Eskinder Ayalew Sisay \n7 (eskinder.ayalew@aau.edu.et)1\n8 1Department of Pharmacology and Clinical Pharmacy, School of Pharmacy, Collage of Health \n9 Sciences, Addis Ababa University, Addis Ababa, Ethiopia.\n10 2Department of Pharmacy, College of Medicine and Health Sciences, Ambo University, Ambo, \n11 Ethiopia.\n12 3Department of Internal Medicine, Addis Ababa University, Addis Ababa, Ethiopia\n13 *Correspondence Author\n14 Hanan Muzeyin Kedir\n15 hanan.muzeyin@aau.edu.et\n16 Department of Pharmacology and Clinical Pharmacy, School of Pharmacy, Collage of Health \n17 Sciences, Addis Ababa University, Addis Ababa, Ethiopia.\n18 Telephone: +251912275828\n19 These authors contributed equally to this work. \n20\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \nNOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.\n\n2\n21 Abstract\n22 Introduction: During hemodialysis (HD), the presence of clots in the dialyzer can diminish \n23 the effective surface area of the device. In severe cases, clot formation in the circuit can halt \n24 treatment and lead to blood loss in the system. Thus, ensuring proper anticoagulation during \n25 HD is crucial to prevent clotting in the circuit while safeguarding the patient from bleeding \n26 risks. This study aimed to evaluate anticoagulation outcomes and related factors in HD patients \n27 with acute kidney injury (AKI) at selected hospitals in Ethiopia.\n28 Method: A prospective, multicenter observational study was carried out between October 1, \n29 2021, and March 31, 2022. The study encompassed all AKI patients undergoing HD at least \n30 once during the study period. Descriptive statistics were utilized to summarize the data, and \n31 multinomial logistic regression analysis was employed to identify factors associated to clotting \n32 and bleeding.\n33 Results: Data were gathered from 1010 HD procedures conducted on 175 patients. \n34 Extracorporeal circuit clotting was detected in 34 patients during 39 (3.9%) dialysis sessions, \n35 while bleeding incidents occurred in 27 patients across 29 (2.9%) sessions. A statistically \n36 significant association was found between both the total number of HD treatments and blood \n37 flow rate with incidents of clotting. Factors such as length of hospitalization, serum creatinine \n38 levels at admission, signs and symptoms associated with uremia, along with utilization of \n39 anticoagulants or antiplatelet medications demonstrated an association with bleeding events.\n40 Conclusion: Clotting affected 19.4% of participants, while bleeding occurred in 15.4%, \n41 underscoring the importance of close monitoring. The frequency of HD sessions and blood \n42 flow rate are correlated with clotting, while hospitalization duration, serum creatinine levels, \n43 uremic symptoms, and anticoagulant use are associated with bleeding events.\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n3\n44 Keywords: Acute kidney injury, Anticoagulation, Hemodialysis, Ethiopia\n45\n46 Introduction \n47 Hemodialysis (HD) constitutes a medical procedure that facilitates the removal of waste \n48 products from the bloodstream externally via a device outfitted with specialized filtration \n49 systems. The predominant cause for arteriovenous access failure is attributed to thrombosis, \n50 representing 80%-85% of cases (1,2). Within the HD protocol, components such as the dialyzer \n51 and other elements of the extracorporeal blood circuit are also vulnerable to thrombotic events. \n52 This includes tubing, arterial and venous bubble traps in addition to needles or catheters utilized \n53 for vascular access—all factors contributing to coagulation phenomena (3,4). Notably, arterial \n54 and venous bubble traps exhibit a heightened susceptibility towards clotting owing to \n55 diminished blood flow rates leading potentially to stasis and hence clot formation. Such \n56 premature cessation of HD sessions precipitates inadequate dialysis alongside possible \n57 hematologic losses. Furthermore clots within this circuitry can herald significant negative \n58 repercussions including reduction in solute clearance during therapy amplified costs, and \n59 increased operational burden (5-7). \n60 Maintaining an optimal balance between insufficient and excessive heparinization is essential \n61 for achieving proper anticoagulation during HD in order to avert bleeding and clot formation \n62 within the extracorporeal circuit, respectively (6, 8). It is imperative to administer the minimal \n63 efficacious dose of anticoagulant to ensure that the dialyzer and venous chamber remain clear \n64 of blood cell fragments while also facilitating rapid hemostasis at the vascular access site \n65 following treatment (5). Although anticoagulant dosing varies among individuals due to patient-\n66 specific factors and HD-related variables, it is generally lower than doses required for full \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n4\n67 anticoagulation. Inadequate levels of anticoagulation may reduce dialysis’s effectiveness in \n68 eliminating waste products from the bloodstream (4-6).\n69 Unfractionated heparin (UFH) is extensively utilized as an anticoagulant during HD owing to \n70 its long-standing history in clinical practice and a desirable safety profile characterized by a \n71 brief half-life (8,9). The European guidelines for optimal HD practices suggest an initial bolus \n72 of 50 IU/kg UFH administered via the arterial access needle, followed by a continuous infusion \n73 maintenance dosage ranging from 500 to 1500 IU of UFH per hour. It's recommended that the \n74 use of UFH be minimized, especially since patients with acute kidney injury AKI often have \n75 compromised health and may require daily dialysis treatments. In instances of clotting or \n76 bleeding, modifications in dosage or the implementation of heparin-free dialysis methods might \n77 become necessary. Although alternative anticoagulation techniques targeting the extracorporeal \n78 circuit exist, their adoption has been limited due to their complex nature and higher requisites \n79 for administration time and workforce involvement for monitoring purposes (10-12).\n80 Critically ill patients often display impaired coagulation and low platelet counts, which \n81 significantly increases their risk of bleeding when systemic anticoagulants are used. As such, it \n82 is advised that these individuals avoid any treatments that further heighten this risk, especially \n83 those involving systemic anticoagulants (6,13). In situations where the risk of bleeding is a \n84 concern, the European Best Practice Guidelines recommend alternatives like regular saline \n85 flushing or local citrate anticoagulation during HD, avoiding systemic anticoagents entirely. \n86 This approach not only facilitates quick adjustments to treatment but also enables easy \n87 monitoring for clotting within the dialyzer and may help in preventing clots from forming \n88 initially (2,10). The goal of this study is to explore outcomes related to anticoagulation therapy \n89 and factors influencing its outcome among AKI patients receiving HD at Tikur Anbessa \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n5\n90 Specialized Hospital (TASH) and St. Paul's Hospital Millennium Medical College (SPHMMC) \n91 in Addis Ababa, Ethiopia.\n92 Material and Methods\n93 Study Setting\n94 The study took place in the dialysis departments of TASH and SPHMMC, located in Addis \n95 Ababa, Ethiopia. TASH introduced HD services back in 1980 and presently caters to acute \n96 dialysis needs. Meanwhile, SPHMMC began its provision of acute dialysis services in August \n97 2013 before extending these services to include maintenance dialysis for kidney transplantation \n98 starting from early 2015 (14).\n99 Despite the lack of a specific written protocol for HD patients, anticoagulation was frequently \n100 administered at both hospitals. UFH was used during HD, but the prescribed doses of heparin \n101 varied slightly between the two centers. At TASH, UFH was administered in one of three ways: \n102 (i) the standard dose, 1 ml (5000 IU) of UFH, was diluted with 4 ml of normal saline (total of \n103 5 ml). This was followed by a 2 ml initial bolus and a continuous infusion of 1 ml/hr; (ii) the \n104 minimal or half dose of 0.5 ml (2500 IU) of UFH was diluted with 4.5 ml of normal saline \n105 (total of 5 ml). This was followed by a 2 ml initial bolus and a continuous infusion of 1 ml/hr; \n106 (iii) heparin-free saline flushes of 30 to 50 ml were given every 30 minutes. At SPHMMC, 1 \n107 ml of UFH was diluted with 3 ml of normal saline (total of 4 ml) and administered as a bolus \n108 of 1 ml, followed by a continuous infusion of 1 ml/hr.\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n6\n109 Study Design and Period\n110 A comprehensive, multicenter observational study was carried out over a duration of six months, \n111 starting from October 1st, 2021 and concluding on March 31st, 2022. This study took place at \n112 two major hospitals (TASH and SPHMMC) in Addis Ababa, Ethiopia. \n113 Sampling and Sample Size Determination\n114 On average, the dialysis clinics at TASH and SPHMMC serve approximately 15 and 30 AKI \n115 patients per month, respectively. All patients meeting the eligibility criteria were considered \n116 for inclusion in the study. Consequently, a total of 175 patients identified during the study \n117 period, meeting the established criteria, participated in this research. Specifically, 122 of these \n118 patients received treatment at SPHMMC, while the remaining 53 were from TASH.\n119 Inclusion and Exclusion Criteria \n120 All patients suffering from AKI, aged 12 years and older, who underwent HD at least once \n121 during the designated study period were eligible for inclusion in this study. Nonetheless, \n122 individuals diagnosed with chronic kidney disease who were undergoing regular maintenance \n123 dialysis treatments, as well as those who expressed a refusal to participate in the study, were \n124 systematically excluded from consideration.\n125 Data Collection Instruments and Techniques\n126 The data collection questionnaire, consisting of four sections, was developed through a \n127 thorough review of relevant literature related to the study's objectives. The aim was to collect \n128 demographic characteristics and clinical variables of patients to assess outcomes related to \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n7\n129 anticoagulation. Pre-testing was carried out, and all necessary adjustments to the data collection \n130 tool were implemented(4,6,9,15–18). \n131 The initial section of the data collection tool was designed to gather information regarding the \n132 sociodemographic characteristics of the patients. Subsequently, the second section assessed the \n133 clinical features of the patients, including inquiries about the cause of AKI, indications for HD, \n134 concomitant medical conditions, co-administered medications, and current use of \n135 anticoagulants and/or antiplatelet drugs. The third part of the data collection tool focused on \n136 recording essential laboratory results upon admission and discharge. Finally, the fourth section \n137 specifically documents the comprehensive details of each HD session. Patients were closely \n138 monitored during each session, with data collectors prospectively recording various parameters \n139 related to HD, such as venous pressure, ultrafiltration volume, blood flow, dialysate flow, and \n140 approach to UFH usage during HD. Additionally, incidents of clotting or bleeding were \n141 documented, along with the interventions implemented to manage these occurrences. \n142 Data Quality Assurance\n143 To ensure data quality, the principal investigator closely monitored each activity to ensure \n144 comprehensive data collection. Unclear terms were promptly clarified, and the necessary \n145 corrections were made during the pre-test phase before commencing the main study. Data \n146 collection was performed using three BSc. nurses and a pharmacist who were recruited as data \n147 collectors. They underwent training to establish a consistent understanding and interpretation \n148 of the instrument, including a briefing on the study's objectives and commitment to maintaining \n149 core ethical principles throughout the data collection period.\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n8\n150 Data Analysis \n151 Data completeness was verified before entering the statistical package for social science \n152 (SPSS) version 26 for a thorough analysis. Descriptive statistics, including the mean, median, \n153 percentage, and standard deviation (SD), were then used to succinctly summarize the data. \n154 Multinomial logistic regression was used to identify predictive factors for anticoagulation-\n155 related outcomes in patients with AKI undergoing HD. A P-value < 0.05 was considered the \n156 threshold for declaring statistically significant associations.\n157 Ethical Considerations\n158 Ethical clearance was secured from Addis Ababa University, School of Pharmacy, the Ethical \n159 Review Committee (reference number ERB/SOP/257/13/2021), and the Institutional Review \n160 Board of SPHMMC (reference number PM25/386). Study participants provided both written \n161 and verbal consent before participating, and their confidentiality was maintained by avoiding \n162 the recording of identifying information such as patient names. The collected data remained \n163 securely stored throughout the study and was appropriately destroyed upon study completion.\n164 Operational Definition\n165 Anticoagulation-related complication: were defined as the occurrence of any circuit clotting or \n166 bleeding during HD, whether with the use of UFH as an anticoagulant for dialysis or in a \n167 heparin-free approach.\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n9\n168 Results\n169 Socio-demographic Characteristics of Study Participants\n170 Of the 175 patients included in the study, an almost equal number were males (50.9%) and \n171 females (49.1%), with a mean age of 40 ± 17.8 years involved and four female were pregnant. \n172 Among the 175 patients, 66.9% were paid out-of-pocket for dialysis and 2.3% were involved \n173 in smoking, drinking alcohol, and chewing khat (Table 1). \n174 Table 1: Socio-demographic characteristics of AKI patients who undergo HD (n=175)\nVariable Category n (%) \nAge (years),mean ± SD 40±17.8\nMale 89 (50.9)Sex\nFemale 86 (49.1)\nSingle 55 (31.4)\nMarried 103 (58.9)\nWidowed 10 (5.7)\nMarital status\nDivorced 7 (4)\nPregnancy Pregnant 4 (4.7)\nNot pregnant 71 (82.5)\nPostpartum 11 (12.8)\nUnable to read and write 23 (13.1)\nPrimary education 42 (24)\nEducation\nSecondary education 63 (36)\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n10\nDiploma and above 47 (26.9)\nWeight (Kg) 61±13.9\nInsurance 58 (33.1)Method of payment\nOut of pocket 117 (66.9)\nSubstance use Smoking 7 (4)\nChewing Khat 9 (5.1)\nAlcohol 26 (14.9)\nSmoking + Khat + Alcohol 4 (2.3)\n175\n176 Clinical Characteristics\n177 The average length of hospital stay was 14.2 days, range, 2–40 days. The three most common \n178 causes of AKI were acute tubular necrosis (29%), nephrotic syndrome (24%), and sepsis \n179 (22.9%). Pregnancy-related causes also included preeclampsia/eclampsia (5.1%) and HELLP \n180 syndrome (4.6%). The major indications for dialysis were uremic signs and symptoms (70.3%), \n181 followed by fluid overload (50.3%). Common underlying comorbidities identified were \n182 hypertension (55.4%) and anaemia (49.7%), and 13.1% of AKI cases were superimposed on \n183 chronic kidney disease. UFH 7500 IU twice per day (48%) was the most widely used \n184 anticoagulant regimen in the study population, followed by 5000 IU twice a day (14.3%) to \n185 prevent deep venous thrombosis (Table 2).\n186 Table 2: Clinical characteristics of AKI patients who undergo HD (n=175)\nn (%)Variables Category\nYes\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n11\nLength of hospitalization (days), mean ± \nSD\n14.2 ± 9.9\nCause of AKI Acute tubular necrosis 49 (28)\nNephritic syndrome 42 (24)\nSepsis 40 (22.9)\nSevere dehydration 17 (9.7)\nInterstitial nephritis 25 (14.3)\nShock 19 (10.9)\nPoison 17 (9.7)\nUrinary tract obstruction 11 (6.3)\nNephrotoxic drugs 8 (4.6)\nRhabdomyolysis 3 (1.7)\nHELLP syndrome 8 (4.6)\nPost-kidney transplant \nrejection\n6 (3.4)\nLupus nephritis 9 (5.1)\nPreeclampsia/Eclampsia 9 (5.1)\nAcute Glomerulonephritis 7 (4)\nHemorrhage 5 (2.9)\nOthers* 4 (2.3)\nDialysis indication Uremic signs and symptoms 123 (70.3)\nFluid overload 88 (50.3)\nHyperkalemia 79 (45.1)\nMetabolic Acidosis 66 (37.7)\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n12\nComorbidity Atrial fibrillation 1 (0.6)\nCoronary arterial disease 1 (0.6)\nStroke 6 (3.4)\nHypertension 97 (55.4)\nCongestive heart disease 22 (12.6)\nGastrointestinal ulceration 29 (16.6)\nChronic Kidney Disease 23 (13.1)\nAnemia 87 (49.7)\nDeep vein thrombosis 5 (2.9)\nPulmonary embolism 3 (1.7)\nType 2 Diabetes 22 (12.6)\nHIV 11 (6.3)\nPulmonary tuberculosis 9 (5.1)\nHepatitis B virus 5 (2.9)\nOthers** 41 (23.4)\nConcurrent anticoagulant and/or \nantiplatelet \n124 (70.9)\nUFH 2500 IU SC BID 1 (0.6)\nUFH 5000IU SC Bid 25 (14.3)\nUFH 7500IU SC Bid 84 (48)\nUFH 17500IU SC Bid 5 (2.9)\nWarfarin 5mg PO Daily 8 (4.6)\nASA 81mg PO Daily 20 (11.4)\nEnoxaparin 40mg SC Bid 3 (1.7)\nClopidogrel 75mg PO Daily 2 (1.1)\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n13\nConcomitant drugs use 166 (94.9)\n187 * (malaria, kidney trauma, cardio-renal syndrome, acute fatty liver of pregnancy)\n188 ** (prostate cancer, BPH, epilepsy, leukemia, gout arthritis, cerebral malaria)\n189 Laboratory Findings\n190 Upon admission and upon discharge, it was noted that there were low hemoglobin levels present. \n191 Furthermore, the mean serum creatinine level recorded at the time of admission stood at 8.3 ± \n192 3.4 mg/dl, which experienced a significant decrease by 3 mg/dl by the time of discharge. Table \n193 3 in our document provides a summarized outline showing these changes among other selected \n194 laboratory parameter values observed during this period.\n195 Table 3: Selected Laboratory Values of AKI patients who undergo HD (n=175)\nMean ± SDVariables\nAdmission Discharge\nHemoglobin (g/dl) 10.3 ± 7.6 10.3 ± 6.4\nWBC count (103/ml) 11.1 ± 5.6 9.8 ± 4.9\nPlatelet count (103/ml) 235.9 ± 140.1 268.3 ± 131\nSerum creatinine (mg/dl) 8.3 ± 3.4 5.3 ± 3.1\nUrea (mg/dl) 137.5 ± 80.3 82.1 ± 63.4\nSerum potassium (mEq/L) 5.3 ± 1.1 4.4 ± 0.8\n196\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n14\n197 Characteristics of HD Sessions\n198 From October 2021 to March 2022, a total of 1010 HD sessions were conducted on 175 patients. \n199 The average frequency of dialysis sessions was calculated at 5.8 ± 3.8 per patient, ranging from \n200 one single session to a maximum of fifteen sessions with each session lasting an average \n201 duration of approximately 3.1 ± 0.6 hours. Patients underwent HD therapy thrice weekly using \n202 catheters as the primary means for vascular access across all cases in this study.  During these \n203 treatments, blood products were administered in approximately  9.8% or essentially, during \n204 almost ten out all hundred HD procedures. UFH was administered for circuit anticoagulation \n205 in 626 sessions, accounting for 62% of the total. The average dose given was 1 mL (5000 IU). \n206 Of these UFH sessions, infusion was stopped 60 minutes before discontinuation in 546 cases, \n207 making up 87.2% of the total. The average pressure in the venous chamber was recorded as \n208 125 ± 39.3 mmHg, with the dialysis blood flow rate averaging 227.2 ± 22.8 ml/min.\n209 Table 4: Characteristics of HD sessions included in the study (N=1010)\nVariables mean ± SD\nTotal number of HD sessions (N=1010) 5.8 ± 3.8\nHD duration (in hour) 3.1 ± 0.6\nVenous chamber pressure (mmHg) 125 ± 39.3\nUltrafiltration volume (ml) 885 ± 650.9\nBlood flow rate (ml/min) 227.2 ± 22.8\nDialysate flow rate (ml/min) 481 ± 46.6\nTotal UFH dose (ml)* 1 ± 0.2\nCategory n (%)\nUFH infusion stopped \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n15\nBefore 30 minutes 70 (11.3)\nBefore 60 minutes 546 (87.2)\nBefore 120 minutes 8 (1.3)\nAt the end of dialysis 1 (0.2)\n210\n211 Anticoagulation-related complication and HD Outcome \n212 Of the 175 patients, extracorporeal circuit clotting occurred in 34 patients in 39 (3.9%) sessions \n213 and 27 patients in 29 (2.9%) experienced bleeding. HD circuit clotting resulted in the \n214 discarding of the bloodline and early termination of dialysis in 17 sessions. Additionally, \n215 nursing interventions, that is, the use of UFH and a normal saline flush, were required in 10 \n216 and 5 sessions, respectively, to prevent abrupt HD treatment interruptions due to circuit \n217 clotting. Some of the measures taken when bleeding occurred included holding UFH at the \n218 time, and for the next session/s (n = 14 sessions), blood transfusion (n = 2 sessions), and \n219 applying pressure at catheter insertion site in 6 sessions (Table 5). On the other hand, the \n220 majority of AKI patients (44.6%) were discharged with improvement, while more than one-\n221 fourth (26.9%) of them died after undergoing HD (Figure 1).\n222 Table 5: Anticoagulation related complication and HD outcome among AKI patients in \n223 the study. \nVariables mean ± SD\nCircuit clotting (34 patients, 19.4%) 39 (3.9)\nMeasures taken when circuit clotting occurs Discard the blood line and \nHD terminated early\n17 (1.7)\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n16\nBlood transfusion during HD 99 (9.8)\nUsing UFH during HD 626 (62)\nAdditional UFH given 10 (1)\nContinued HD by using \nnormal saline flush\n5 (0.5)\nReturn the blood and HD \nterminated early\n2 (0.2)\nManipulate the access and \ncontinued HD\n3 (0.3)\nUse UFH intermittently \nfor the next sessions\n2 (0.2)\nBleeding (27 patients, 15.4%) 29 (2.9)\nBleeding from nose 15 (1.5)\nCatheter site bleeding 6 (0.6)\nVaginal bleeding 4 (0.4)\nOthers** 4 (0.4)\nMeasures taken when bleeding occurs Hold UFH at the time, and  \nfor the next session/s\n14 (1.4)\nBlood transfused 2 (0.2)\nApplied pressure at \ncatheter insertion site\n6 (0.6)\nUse UFH intermittently \nfor the next sessions\n2 (0.2)\nReturn the blood and HD \nterminated early\n1 (0.1)\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n17\nNo measure taken 4 (0.4)\n224 *1ml=5000 IU\n225 ** bleeding form rectum, bloody vomiting, subcutaneous hematoma, bleeding form bullet \n226 penetrating injury site\n227 Factors associated with anticoagulation related outcomes\n228 To determine the predictors of anticoagulation-related outcomes during HD (clotting and \n229 bleeding risks), multinomial regression analysis was used. Based on the model comparing \n230 participants who had HD circuit clotting with those who did not have any clotting/bleeding, \n231 as the total number of HD sessions increased, the probability of circuit clotting increased \n232 (AOR=1.932, 95% CI, 1.227-3.043, p= 0.004). In contrast, a higher blood flow rate was \n233 associated with a lower HD circuit clotting risk (AOR=0.868, 95% CI, 0.812-0.928, p= \n234 0.001). \n235 On the other hand, bleeding was more likely to occur in participants who had longer \n236 hospitalization and elevated serum creatinine at admission (AOR=1.247, 95% CI, 1.053-1.478, \n237 p= 0.010; AOR=1.886, 95% CI, 1.285-2.769, p= 0.001) respectively. While, the rate of \n238 bleeding in patients who had no uremic signs and symptoms was lower (AOR=0.092, 95% CI, \n239 0.009-0.955, p= 0.004) than those having it. Likewise, patients who did not receive \n240 prophylactic or therapeutic anticoagulant and/or antiplatelet drug were less likely to develop \n241 bleeding (AOR=0.017, 95% CI, 0.001-0.446, p= 0.014) compared to those taking these drugs \n242 (Table 6).\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n18\n243 Table 5: Multinomial regression of predictive factors associated with anticoagulation \n244 related outcomes during HD\nOutcomesVariables\nClotting\nAOR (95% CI)\nBleeding\nAOR (95% CI)\nP-value\nTotal no of HD sessions 1.932 (1.227-3.043)a 1.068 (0.713-1.601) 0.004*\nBlood flow rate (ml/min) 0.868 (0.812-0.928)a 1.026 (0.974-1.081) <0.001*\nLength of hospitalization \n(days)\n0.946 (0.843-1.06) 1.247 (1.053-1.478)a 0.010*\nSerum creatinine (mg/dl) at \nadmission\n0.911 (0.699-1.189) 1.886 (1.285-2.769)a 0.001*\nUremic signs and symptoms\nYes 1 1\nNo 0.532 (0.079-3.57) 0.092 (0.009-0.955)a 0.04*\nUsing of anticoagulant \nand/or antiplatelet drug\nYes 1 1\nNo 1.156 (0.168-7.936) 0.017 (0.001-0.446)a 0.014*\n245 Reference Variable: No clotting/bleeding\n246 *Variables that showed a significant association, P < 0.05\n247 aVariable category with P <0.05\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n19\n248 Discussion\n249 As the first prospective study in the Ethiopian context on anticoagulation among patients with \n250 AKI on HD, this study provides insight into extracorporeal circuit clotting or bleeding \n251 occurrences and the predictive factors associated with these complications. In this study, 1010 \n252 HD sessions performed on 175 patients were prospectively followed. Clotting and bleeding \n253 data are expressed as percentages of both participants and sessions. Circuit clotting occurred \n254 in 3.9% of HD sessions (19.4% patients), and bleeding in 2.9% sessions (15.4% patients). \n255 A similar result was reported in the United States of America, that evaluated clotting of the HD \n256 circuit in general care patients and critically ill patients. The overall rate of extracorporeal \n257 circuit clotting was 5.2% in all sessions (12). Circuit clotting occurred in 2.4% of HD sessions \n258 with heparin, according to a retrospective study comparing two intra-dialytic heparin protocols: \n259 \"routine heparin-use\" during HD (routine heparin prime/bolus dose) and heparin-free HD \n260 (saline prime, heparin avoidance). However, heparin-free HD is associated with a higher rate \n261 of HD circuit clotting (9.1%) (13). In contrast, other studies reported a higher rate of circuit \n262 clotting during HD. A study conducted in Brazil showed that filter clotting occurred in 19 \n263 patients (25.3%) and 29 sessions (14.9%) (19). This study evaluated and compared intra- and \n264 post-dialysis complications in critically ill AKI patients undergoing extended daily dialysis \n265 sessions of different durations (6 versus 10 hours), which could explain the higher rate of HD \n266 filter clotting events, and the fact that many critically ill patients develop hemostatic \n267 abnormalities in addition to the longer dialysis duration.\n268 Additionally, a retrospective cohort study conducted in Belgium found that circuit coagulation \n269 was reported in 17.5% of the sessions. This higher rate of clotting events could be attributed to \n270 the absence of systemic anticoagulation (20). Moreover, this study conducted a retrospective \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n20\n271 analysis of HD sessions in intensive care unit patients, which could be an additional factor \n272 contributing to clotting occurrences. Patients admitted to the intensive care unit and requiring \n273 dialysis for AKI often present with a systemic inflammatory state (21), which is known to be \n274 associated with the activation of coagulation pathways (22). \n275 Lower clotting of the HD circuit incidence was reported in a retrospective study conducted in \n276 the USA (1% of 400 HD treatments) (23). This may be attributed to the operational definition \n277 of clotting used in this study. The researchers defined clotting as complete clotting that required \n278 replacement of the blood tubing and dialyzer to complete treatment. The higher blood flow rate \n279 (378 ± 46 mL/min) and more aggressive normal saline flushing of the circuit employed in this \n280 protocol could also explain the lower rate of clotting. Moreover, there might be underreporting \n281 of clots due to the retrospective study design, in which existing data were extracted from chart \n282 reviews. Another study from Spain found that 1.9% of patients developed clotting, which is \n283 lower than our finding (19.4% of patients) with a higher mean blood pump flow (346 ± 47 \n284 ml/min), given a higher risk of coagulation from a lower pump flow (17).\n285 In the current study, 15.4 % of patients developed bleeding, and this finding was higher than a \n286 study conducted in Spain (4.4%) (17). This variation might be due to the difference in the study \n287 design and setting, even though the study reported that oral anticoagulants were administered \n288 to 18.4% of patients, which by itself is thought to increase the risk of bleeding. However, in a \n289 Spanish study, each patient was asked to report any bleeding or thrombotic complications that \n290 arose in the week prior to data collection.\n291 In the present study, it was observed that anticoagulation was achieved with UFH or saline \n292 flushes of 30 to 50 ml administered every 30 min if UFH was contraindicated. There were three \n293 strategies regarding UFH order for HD in current study settings, and each patient was dialyzed \n294 using any of these anticoagulation methods. The first is the standard dosing regimen of UFH, \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n21\n295 consisting of a bolus of 1250–2000 U, followed by a continuous infusion of 1000–1250 U/h. \n296 The second was the minimal/half-dose method with a 625–1000 bolus dose and 500–625 \n297 infusion rate. Heparin-free dialysis was also the third method used in patients who were \n298 actively bleeding or were at an increased risk of bleeding. \n299 Due to its brief half-life and extensive medical usage over time, UFH has emerged as the \n300 predominant anticoagulant for HD (9). Nonetheless, the dosage administered shows significant \n301 divergence. This study reveals that the average total UFH dose throughout HD sessions was \n302 5000 ± 1000 units, with heparin infusion being discontinued one hour before dialysis \n303 completion in the majority (87.2%) of cases to reduce the potential for bleeding from the access \n304 site post needle withdrawal. \n305 In the present study, extracorporeal circuit clotting was assessed by visual inspection and noted \n306 by nurses. It is characterized by extremely dark blood or black striations in the tube and a \n307 sudden rise in pressure readings. They only recorded clotting as being present or absent without \n308 grading it in line with several studies done elsewhere (13,17,19,24). In contrast, other studies \n309 have reported the degree of clotting by classifying it as mild/slight, moderate, and severe.\n310 HD is often performed three times a week for three–four hours each session; however, the \n311 length of these sessions varies from patient to patient. One consequence of intradialysis, circuit \n312 clotting, includes treatment interruption and patient blood loss, which may exacerbate \n313 haemodynamic instability (19). In 1.7% of all sessions with clotting, dialysis was stopped early \n314 without the possibility of retransfusing blood from the extracorporeal circuit; hence, the \n315 bloodline was discarded. However, in 0.2% of the sessions, dialysis was terminated early by \n316 the return of blood.\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n22\n317 According to a study from Belgium, clotting reduced the length of time spent receiving dialysis \n318 in 15.2% of sessions, completely blocking the circuit, and preventing retransfusion in 4.2% of \n319 sessions, which is greater than our findings (20). These variations may be due to differences in \n320 dialysis protocols. In contrast to our study, the study combined a citrate-enriched dialysate with \n321 a heparin-coated dialyzer with no systemic UFH. Similarly, a study conducted in the USA \n322 comparing two intradialytic heparin protocols reported that in a heparin-free HD protocol, \n323 circuit clotting resulted in a change of the extracorporeal circuit in 7.3% of sessions. However, \n324 this was required in only 0.8% of HD sessions with heparin, and early termination of HD \n325 (1.6%) was similar to that in this study (13). Bleeding is another important intradialysis \n326 complication that might require adjusting the dose of UFH (1.6% of all sessions with bleeding) \n327 or blood transfusion (0.2%). Another study revealed that patient weight, circuit clotting, and \n328 bleeding of the vascular access after disconnection were the most frequently employed \n329 techniques for changing the dosage (17). \n330 Identifying predictors of anticoagulation-related outcomes during HD in patients with AKI is \n331 important. Identifying and preventing patients at risk of such complications can increase \n332 practitioner awareness of HD-related complications. Similar to other studies (12,13,20,23), the \n333 blood flow rate showed a statistically significant association with the occurrence of circuit \n334 clotting. In this study, lower delivered blood flow rates were associated with higher circuit-\n335 clotting rates. Additionally, as the total number of HD sessions increased, the probability of \n336 circuit clotting increased, unlike in a study conducted in Belgium (20). This difference could \n337 be due to variations in their research methodology and the inclusion of study participants with \n338 varying levels of clotting risk. Patients who did not receive prophylactic or therapeutic \n339 anticoagulants and/or antiplatelet drugs were less likely to develop bleeding compared to those \n340 taking these drugs. Similarly, a study from Spain found that bleeding complications were more \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n23\n341 frequent in patients receiving oral anticoagulants (17). Consistent with previous studies, age \n342 and sex were not associated with bleeding events (13,17).\n343 However, there were some limitations to our study. Bleeding was reported only at the time of \n344 the dialysis session. Therefore, this may not provide a comprehensive picture of the bleeding \n345 risk associated with dialysis. Additionally, the study did not exclude patients with \n346 comorbidities (such as deep vein thrombosis, active malignancy, severe heart failure, or severe \n347 liver disease). These conditions can impact anticoagulation outcomes and may confound the \n348 interpretation of results.\n349 Conclusions\n350 In conclusion, clotting was observed in 19.4% of the participants in the study, while bleeding \n351 was reported in 15.4% of them. These findings highlight the importance of thorough \n352 monitoring due to the complications resulting in treatment interruptions and blood loss. The \n353 study identified a significant association between circuit clotting and variables such as blood \n354 flow rate and the frequency of HD sessions. Furthermore, it revealed a notable occurrence of \n355 bleeding during HD, mainly manifesting as minor cases. Individuals who received prophylactic \n356 or therapeutic anticoagulants and/or antiplatelet medications were more susceptible to bleeding \n357 compared to those who did not undergo these treatments.\n358\n359 Abbreviations\n360 AKI: Acute Kidney Injury; HD: Hemodialysis; SPHMMC: St. Paul's Hospital Millennium \n361 Medical College; TASH: Tikur Anbessa Specialized Hospital; UFH: Unfractionated heparin.\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n24\n362 Acknowledgement\n363 The authors thank the School of Pharmacy, College of Health Sciences, Addis Ababa \n364 University, and the nursing staff at TASH and SPHMMC dialysis clinics, who were very \n365 cooperative and helpful in providing constant assistance.\n366 Availability of Data and Materials\n367 All the data and materials used in this paper are available from the corresponding author upon \n368 reasonable request.\n369 Disclosure interest\n370 The authors declare that they have no conflicts of interest.\n371 Funding details\n372 The authors received no specific funding for this study.\n373\n374\n375\n376\n377\n378\n379\n380\n381\n382\n383\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n25\n384 References\n385 1. Chawla LS, Amdur RL, Amodeo S, Kimmel PL, Palant CE. The severity of acute \n386 kidney injury predicts progression to chronic kidney disease. Kidney Int. \n387 2011;79(12):1361–9. \n388 2. Daugirdas JT, Blake PG, Ing TS. Handbook of dialysis: Fifth edition. Handbook of \n389 Dialysis: Fifth Edition. 2014. 1–900 p. \n390 3. Dirkes S, Wonnacott R. Continuous Renal Replacement Therapy and Anticoagulation: \n391 What Are the Options? Am Assoc Crit Nurses. 2016;36(2):1–12. \n392 4. Ross S. Anticoagulation in Intermittent Hemodialysis: Pathways, Protocols, and \n393 Pitfalls. Vet Clin North Am - Small Anim Pract. 2011;41(1):163–75. \n394 5. Fischer KG. Essentials of anticoagulation in hemodialysis. Hemodial Int. \n395 2007;11(2):178–89. \n396 6. Mhatre V. Ho, Ji-Ann Lee  and KCM. Use and Safety of Unfractionated Heparin for \n397 Anticoagulation During Maintenance Hemodialysis. Am J Kidney Dis. 2012;23(1):1–\n398 7. \n399 7. Srisawat N, Lawsin L, Uchino S, Bellomo R, Kellum JA. Cost of acute renal \n400 replacement therapy in the intensive care unit: Results from The Beginning and Ending \n401 Supportive Therapy for the Kidney (BEST Kidney) Study. Crit Care. 2010;14(2):1–10. \n402 8. Kessler M, Moureau F, Nguyen P. Anticoagulation in Chronic Hemodialysis: Progress \n403 Toward an Optimal Approach. Semin Dial. 2015;28(5):474–89. \n404 9. Cronin RE, Reilly RF. Unfractionated Heparin for Hemodialysis: Still the Best Option. \n405 Semin Dial. 2010;23(5):510–5. \n406 10. European Best Practice Guidelines Expert Group on Hemodialysis ERA. Section V. \n407 Chronic intermittent haemodialysis and prevention of clotting in the extracorporal \n408 system. Nephrol Dial Transpl. 2002;17:63–71. \n409 11. Royal Liverpool and Broadgreen University Hospitals NHS Trust. Protocol for \n410 anticoagulation of extracorporeal circuits during kidney replacement therapy or plasma \n411 exchange. 2015. \n412 12. Safadi S, Albright RC, Dillon JJ, Williams AW, Alahdab F, Brown JK, et al. \n413 Prospective Study of Routine Heparin Avoidance Hemodialysis in a Tertiary Acute \n414 Care Inpatient Practice. Kidney Int Reports [Internet]. 2017;2(4):695–704. Available \n415 from: http://dx.doi.org/10.1016/j.ekir.2017.03.003\n416 13. Liang E. Outcomes Associated with a Heparin-Free Hemodialysis Protocol and \n417 Review of the Literature. J Clin Nephrol Ren Care. 2016;2(1):1–6. \n418 14. Muleta MB, Abebe E, Tadesse M, Berhae T, Muhammed M, Woodside K, et al. \n419 Original Article Milestones of Renal Replacement Therapy in Ethiopia. Ethiop Med J. \n420 2020;Supplement(January):5–13. \n421 15. Chan KE, Lazarus JM, Thadhani R, Hakim RM. Anticoagulant and antiplatelet usage \n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n26\n422 associates with mortality among hemodialysis patients. J Am Soc Nephrol. \n423 2009;20(4):872–81. \n424 16. Krummel T, Scheidt E, Borni-duval C, Bazin D, Lefebvre F, Nguyen P, et al. \n425 Haemodialysis in patients treated with oral anticoagulant : should we heparinize ? \n426 2014;29(4):906–13. \n427 17. Herrero-Calvo JA, González-Parra E, Pérez-García R, Tornero-Molina F. Spanish \n428 study of anticoagulation in haemodialysis. Rev Nefrol. 2012;32(2):153–9. \n429 18. Shibiru T, Gudina EK, Habte B, Deribew A, Agonafer T. Survival patterns of patients \n430 on maintenance hemodialysis for end stage renal disease in Ethiopia : summary of 91 \n431 cases. BMC Nephrol. 2013;14(1–6). \n432 19. Albino BB, Balbi AL, Ponce D. Dialysis complications in AKI patients treated with \n433 extended daily dialysis: Is the duration of therapy important? Biomed Res Int. \n434 2014;2014. \n435 20. François K, Wissing KM, Jacobs R, Boone D, Jacobs K, Tielemans C. Avoidance of \n436 systemic anticoagulation during intermittent haemodialysis with heparin-grafted \n437 polyacrilonitrile membrane and citrate-enriched dialysate: A retrospective cohort \n438 study. BMC Nephrol. 2014;15(1):1–7. \n439 21. Aleksandrova I V., Reǐ SI, Pervakova EI. Acute renal failure in critically ill patients. \n440 Anesteziol Reanimatol. 2007;294(4):72–6. \n441 22. Franchini M, Veneri D, Lippi G. Inflammation and hemostasis: A bidirectional \n442 interaction. Clin Lab. 2007;53(1–2):63–7. \n443 23. Sahota S, Rodby R. Inpatient hemodialysis without anticoagulation in adults. Clin \n444 Kidney J. 2014;7(6):1–5. \n445 24. Murea M, Russell GB, Daeihagh P, Saran AM, Pandya K, Cabrera M, et al. Efficacy \n446 and safety of low-dose heparin in hemodialysis. Hemodial Int. 2018;22(1):74–81. \n447\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint \n\n . CC-BY 4.0 International licenseIt is made available under a \nperpetuity. \n is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint \nThe copyright holder for thisthis version posted February 29, 2024. ; https://doi.org/10.1101/2024.02.28.24303493doi: medRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}