5. Results
5.2 AKI among patients undergoing hip or knee joint replacement (Study 1)
5.2.2 Risk Factors of AKI
Univariable analysis showed age, BMI, ASA classification, SCr, eGFR and hemoglobin to be associated with postoperative AKI (Table 11).
54
Table 9. Demographic data of Study 4
Variable n /Median %/IQR
All joint replacements 18,979 100%
Age, years 68 61-75
Preoperative SCr 70 61-82
Preoperative eGFR (CKD EPI),
Hypercholesterolemia 4,642 24.6%
Cardiovascular disease 2,474 13.1%
CD = Coronary artery disease, PCI = Percutaneous coronary intervention, CABG = Coronary artery bypass grafting, ACS = Acute coronary syndrome, CHF = Congestive heart failure, Pulmonary disease including asthma, copd, sleep apnea, lung operations, chronic bronchitis and other pulmonary diseases
55
Table 10. Perioperative data of Study 4
Variable n /Median % / IQR
All joint replacements 18,979 100%
Hip replacement 7,753 40.9%
Cementless fixation 4,458 57.5%
Hybrid fixation 2,325 30.0%
Cemented fixation 970 12.5%
Knee replacement 11,226 59.1%
Cementless fixation 857 7.6%
Hybrid fixation 237 2.1%
Cemented fixation 10,132 90.3%
Bilateral knee replacement 859 7.7%
Bilateral hip replacement 244 3.1%
Anesthesia modality (missing 1,389)
Spinal/epidural anesthesia 17,190 98.7%
General anesthesia 218 1.3%
Duration of operation, minutes 84 68-107
5.2 AKI among patients undergoing hip or knee joint replacement (Study 1)
5.2.1 Prevalence of AKI (Study 1)
There were 44 AKI cases among 13,214 (incidence 3.3/1000 operations; 95% CI 2.5-4.5) patients undergoing surgery and included in our hospital laboratory register in their home municipality. Laboratory follow-up data was thus also available after discharge from the joint replacement hospital. Fifty-eight AKI cases were identified among the 18,575 patients. Thus, 14 of the AKI patients had different laboratory registers in their home municipalities and therefore their laboratory test results were lacking after discharge from the joint replacement hospital. None of the patients diagnosed with AKI had dialysis preoperatively.
5.2.2 Risk Factors of AKI (Study 1)
Univariable analysis showed age, BMI, ASA classification, SCr, eGFR and hemoglobin to be associated with postoperative AKI (Table 11).
56
Table 11. Association of preoperative factors and AKI, univariable regression results
Descriptor AKI† non- AKI incidence
/1000
Preoperative SCr 78 (49-150) 70 (25-1125) 1.004 (1.00-1.01) 0.007
Preoperative eGFR
† In this column, numbers represent number of patients unless otherwise stated ref = Reference category
1 ASA classes 1 and 2 were combined for the regression analysis.
2 eGFR calculated using CKD-EPI formula
3Anemia was defined as hemoglobin <117 g/l in women and <134 g/l in men.
57
Univariable analysis showed that revision joint replacement and intravenous antibiotic prophylaxis were associated with postoperative AKI (Table 12). Practically all cement used was impregnated with aminoglycosides such as gentamicin and tobramycin.
Multivariable analysis showed that independent risk factors for AKI were duration of operation, ASA classification, BMI and preoperative eGFR (Table 13). Although all
Table 12. Association of perioperative factors and AKI, univariable regression results
Descriptor AKI† non- AKI incidence
/1000
† In this column, numbers represent number of patients unless otherwise stated ref = Reference category
1 Duration in 10-minute periods from incision to end of wound closure
56
Table 11. Association of preoperative factors and AKI, univariable regression results
Descriptor AKI† non- AKI incidence
/1000
Preoperative SCr 78 (49-150) 70 (25-1125) 1.004 (1.00-1.01) 0.007
Preoperative eGFR
† In this column, numbers represent number of patients unless otherwise stated ref = Reference category
1 ASA classes 1 and 2 were combined for the regression analysis.
2 eGFR calculated using CKD-EPI formula
3Anemia was defined as hemoglobin <117 g/l in women and <134 g/l in men.
57
Univariable analysis showed that revision joint replacement and intravenous antibiotic prophylaxis were associated with postoperative AKI (Table 12). Practically all cement used was impregnated with aminoglycosides such as gentamicin and tobramycin.
Multivariable analysis showed that independent risk factors for AKI were duration of operation, ASA classification, BMI and preoperative eGFR (Table 13). Although all
Table 12. Association of perioperative factors and AKI, univariable regression results
Descriptor AKI† non- AKI incidence
/1000
† In this column, numbers represent number of patients unless otherwise stated ref = Reference category
1 Duration in 10-minute periods from incision to end of wound closure
58
cement used was impregnated with aminoglycosides, use of bone cement was not associated with AKI.
First, a sensitivity analysis was performed including only primary joint replacement operations (n = 15,943). This multivariable analysis identified the same independent risk factors.
Second, a sensitivity analysis was performed including only variables shown in DAG, including BMI, fixation method, bilateral operation, and operation type (primary, revision) and removing ASA classification and operation duration from the multivariable model. The result remained unchanged.
Table 13. Risk factors of AKI, multivariable regression results
Descriptor OR (95% CI) p value
Duration of operation1 1.11 (1.03-1.17) 0.003 ASA classification
1 and 2 Ref
3 4.43 (1.76-11.16) 0.002
4 13.08 (3.71-46.16) <0.001
Bilaterality
Unilateral Ref
Bilateral 0.31 (0.06-1.59) 0.161
Body Mass Index 1.07 (1.01-1.12) 0.015 Fixation technique
Cementless Ref
Hybrid 1.70 (0.56-5.14) 0.348
Cemented 1.19 (0.39-3.67) 0.757
Joint
Hip Ref
Knee 1.03 (0.43-2.46) 0.956
Operation type
Primary Ref
Revision 0.46 (0.12-1.70) 0.245 Preoperative eGFR2 0.98 (0.97-1.00) 0.030
1 Duration in 10-minute periods time from incision to end of wound closure
2eGFR calculated using CKD-EPI formula
59
Third, a sensitivity analysis was performed excluding patients whose data was now on some other laboratory register in their home municipality, and for whom there was therefore no follow up data available after discharge from our unit. (n = 5,361). This analysis showed the same independent risk factors for AKI as did the original multivariable analysis.
Incidence of AKI was higher when operation duration exceeded the arbitrary boundary of two hours (Figure 3).
Figure 3. Incidence of AKI in different operation durations 1.82
2.70
4.89 4.57
0 1 2 3 4 5 6
AKI incidence /1000 operations
Duration of operation
under 90min 90-120min 120-150min over 150min
59
Third, a sensitivity analysis was performed excluding patients whose data was now on some other laboratory register in their home municipality, and for whom there was therefore no follow up data available after discharge from our unit. (n = 5,361). This analysis showed the same independent risk factors for AKI as did the original multivariable analysis.
Incidence of AKI was higher when operation duration exceeded the arbitrary boundary of two hours (Figure 3).
)LJXUH Incidence of AKI in different operation durations 1.82
2.70
4.89 4.57
0 1 2 3 4 5 6
AKI incidence /1000 operations
Duration of operation
under 90min 90-120min 120-150min over 150min
60