- Email: [email protected]
Association of oliguria with the development of acute kidney injury in the critically ill
http://www.kidney-international.org
clinical investigation
© 2015 International Society of Nephrology
Association of oliguria with the development of acute kidney injury in the critically ill Suvi T. Vaara1, Ilkka Parviainen2, Ville Pettilä1, Sara Nisula1, Outi Inkinen3, Ari Uusaro2 and The FINNAKI Study Group4 1
Intensive Care Medicine, Department of Perioperative, Intensive Care and Pain Medicine, Helsinki University Central Hospital, Helsinki, Finland; 2Division of Intensive Care, Kuopio University Hospital, Kuopio, Finland and 3Department of Intensive Care, Turku University Hospital, Turku, Finland
Urine output (UO) criterion may increase the sensitivity of the definition of acute kidney injury (AKI). We determined whether the empirically derived definition for oliguria (o0.5 ml/kg/h) is independently associated with adverse outcome. Data analysis included hourly recorded UO from the prospective, multicenter FINNAKI study conducted in 16 Finnish intensive care units. Confounder-adjusted association of oliguria of different severity and duration primarily with the development of AKI defined by creatinine criterion (Cr-AKI) or renal replacement therapy (RRT) was assessed. Secondarily, we determined the association of oliguria with 90-day mortality. Of the 1966 patients analyzed for the development of AKI, 454 (23.1%) reached this endpoint. Within this AKI cohort, 312 (68.7%) developed Cr-AKI, 21 (4.6%) commenced RRT without Cr-AKI, and 121 (26.7%) commenced RRT with Cr-AKI. Episodes of severe oliguria (o0.1 ml/kg/h) for more than 3 h were independently associated with the development of Cr-AKI or RRT. The shortest periods of consecutive oliguria independently associated with an increased risk for 90-day mortality were 6–12 h of oliguria from 0.3 to o0.5 ml/kg/h, over 6 h of oliguria from 0.1 to o0.3 ml/kg/h, and severe oliguria lasting over 3 h. Thus, our findings underlie the importance of hourly UO measurements. Kidney International advance online publication, 9 September 2015; doi:10.1038/ki.2015.269
Correspondence: Suvi T. Vaara, Intensive Care Medicine, Department of Perioperative, Intensive Care and Pain Medicine, Helsinki University Central Hospital, Box 340, 00290 Helsinki, Finland. E-mail: [email protected] 4
See appendix.
Received 27 October 2014; revised 3 July 2015; accepted 9 July 2015 Kidney International
Acute kidney injury (AKI) is a frequently diagnosed syndrome in the intensive care unit (ICU) that associates with increased mortality.1–3 The diagnosis of AKI is based on an abrupt increase in plasma creatinine (Cr) and/or a decrease in urine output (UO).4–6 Even small increases in Cr are associated with increased short- and long-term mortality and morbidity.2,6 Cr criteria for AKI have been extensively studied7 and recently updated.6 On contrary, the UO criterion has been empirically derived4,5 and its further validation has been suggested.5 Few studies, however, have been able to employ it without modifications. Combining UO and Cr criteria increases the sensitivity of the AKI definition.3,8–10 A recent database analysis showed that patients with both UO and Cr criteria had highest short- and long-term mortality rates.11 Isolated oliguria without fulfillment of Cr criterion also associated with long-term mortality11 and in some studies UO has outperformed Cr in mortality prediction.12,13 Oliguria (UO o0.5 ml/kg/h) has been shown to associate with occurrence of AKI by Risk, Injury, Failure, Loss of Function, and End-stage Renal Disease Cr criterion, but to be only a fair predictor of consequent AKI.14 Among patients with septic shock, consecutive oliguria of 3–5 h has been demonstrated to predict occurrence of Kidney Disease: Improving Global Outcomes (KDIGO) stage 2 or 3 AKI by Cr criterion.15 However, a single center study found that UO o0.3 ml/kg/h for 6 h corresponded better the current Cr-based definition of AKI than UO o0.5 ml/kg/h for 6 h, and was associated with mortality after adjustments whereas UO o0.5 ml/kg/h was not.16 Moreover, UO is prone to confounders such as use of diuretics, differences in fluid management and balance, and in administration of renal replacement therapy (RRT). In this large prospective multicenter study we aimed to find out whether the empirically derived threshold for oliguria o0.5 ml/kg/h is independently associated with adverse outcome. We studied the confounder-adjusted association of oliguria of different severity (mild oliguria from 0.3 to o0.5 ml/kg/h, moderate oliguria from 0.1 to o0.3 ml/kg/h, and severe oliguria o0.1 ml/kg/h) and their duration (3–6 h, 6–12 h, 12–24 h, and over 24 h) 1
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
2901 Patients in the FINNAKI study Exclusions: - 560 With incomplete UO recording within first 24 h - 6 Without baseline weight - 11 With renal transplant - 9 With acute RRT already before ICU admission - 96 With UO recordings done using different method - 63 With incomplete data - 188 Readmissions
2160 Patients analyzed regarding 90-day mortality -90-day mortality 476 (22.0%; 95% CI 20.3–23.8%) 194 (8.9%) Excluded with Cr-AKI/ RRT within 3 h of ICU admission
1966 Patients analyzed regarding Cr-AKI/RRT endpoint -endpoint positive in 454 (23.1%; 95% CI 21.2–25.0%)
Figure 1 | Study flow-chart with excluded patient groups. CI, confidence interval; Cr-AKI, acute kidney injury by creatinine criterion; ICU, intensive care unit; RRT, renal replacement therapy; UO, urine output.
primarily with the development of AKI defined by Cr criterion (Cr-AKI) or administration of RRT. Secondarily, we studied the association of oliguria with 90-day mortality. RESULTS
After exclusions presented in the study flow chart (Figure 1), 2160 patients were included in the analysis. Altogether 924 patients (42.8%) fulfilled the KDIGO criteria (UO, Cr, or RRT) for AKI. Of all, 225 (10.4%) patients received RRT, and 222 patients commenced RRT within the first 24 h in the ICU. The 90-day mortality (95% confidence interval) of all AKI patients was 29.3% (26.4–32.3%). Patients, who fulfilled both Cr and UO criteria, were most severely ill and had the highest rates for administration of RRT and 90-day mortality (Table 1). Oliguria
At least one episode of oliguria (UO below 0.5 ml/kg/h for a minimum of 0.5 h) occurred in 1990 (92.1%) patients. Episodes longer than 30 min of moderate oliguria (UO o0.3 ml/kg/h) occurred in 1694 (78.4%) of the patients and of severe oliguria (UO o0.1 ml/kg/h) in 900 (41.7%) of patients. The median (interquartile range) day of the beginning of the longest oliguric episode was day 1 (0–2). Altogether 608 (28.1%) patients had oliguria with UO o0.5 ml/kg/h for ⩾ 6 consecutive hours, and thus, oliguric AKI. Of these, 330 (54.3%) had stage 1 AKI, 132 (21.7%) stage 2 AKI, and 146 (24.0%) stage 3 AKI according to the KDIGO UO criterion. As short episodes of oliguria of all severities were very frequent, categories of UO from 0.3 to o0.5 ml/kg/h and from 0.1 to o0.3 ml/kg/h were formed by excluding patients with more severe oliguria for longer than 3 h. The incidence of oliguric episodes (the longest episode of each patient included) of different degrees of severity is presented in Table 2. 2
Development of AKI defined by Cr criterion and/or administration of RRT
We studied the association of oliguria and development of AKI based on Cr criterion and/or initiation of RRT (AKI by Cr/RRT) in 1966 patients. Because 194 patients developed AKI by Cr/RRT within 3 h from ICU admission, they were excluded from the analysis regarding the development of AKI. UO data were screened until the diagnosis of AKI by Cr/RRT and among those who did not develop AKI, until ICU discharge or day 5. Of the 1966 patients, 454 (23.1%) developed AKI by Cr/RRT, 312 (68.7%) of them developed Cr-AKI, 21 (4.6%) commenced RRT without Cr-AKI, and 121 (26.7%) commenced RRT with Cr-AKI. The median (interquartile range) time to reach AKI diagnosis by Cr/RRT was 16.9 (10.7–34.3) hours. Altogether 1747 (88.9%) patients had at least one oliguric episode lasting over 30 min. Oliguric AKI was diagnosed in 147 (32.4%) patients before the diagnosis of AKI by Cr/RRT and in 276 (18.3%) of those who did not develop AKI by Cr/RRT criteria (Po0.001). Episodes of UO o0.5 ml/kg/h for more than 12 h occurred in 94 (4.8%) patients, and 52 (55.3%) of them developed AKI by Cr/RRT. None of the patients presented with oliguria lasting over 24 h. Figure 2 presents the proportions of patients who developed AKI by Cr/RRT according to the severity and duration of oliguria. Of the 1042 patients with UO o0.5 ml/ kg/h for more than 3 h, only 248 (23.8%) did not receive diuretics in the ICU. We constructed logistic regression models to study the independent association of different types of oliguric episodes with the development of AKI by Cr/RRT. The reference group comprised those with no oliguria or oliguric episodes lasting o3 h. We adjusted for age, sex, Acute Physiology and Chronic Health Evaluation (APACHE) II diagnosis group, Simplified Kidney International
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
Table 1 | Incidence, patient characteristics, and outcomes according to the presence of UO and/or Cr criteria for AKI No of patients – number (%; 95% CI) Age Male sex Chronic kidney disease Chronic heart failure Hypertension Operative admission Emergency admission SAPS II score Non-renal non-age SAPS II score SOFA score, day 1 Severe sepsis diagnosed by day 5 Use of vasoactives (admission day) Use of diuretics (admission day) Cumulative UO within first 24 h in ICU (ml/kg/h) Fluid accumulation % of baseline weight on day 1
No AKI
UO-AKI
Cr-AKI
Cr and UO AKI
P-valuea
1246 (57.7; 55.6–59.8%) 63 (52–73) 795 (63.8) 43 (3.5) 138 (11.1) 548 (44.0) 495 (39.7) 1036 (83.1) 34.0 (27.0–44.0) 21.0 (14.0–30.0) 7 (4–8) 317 (25.4) 732 (58.7) 357 (28.7) 2500 (1859–3546) 1.4 (−1.1–3.7)
296 (13.7; 12.3–15.2%) 64 (54–73) 196 (66.2) 21 (7.1) 35 (11.8) 146 (49.3) 123 (41.6) 247 (83.4) 36.5 (29.0–50.0) 23.0 (16.0–33.0) 7 (6–9) 82 (27.7) 221 (74.7) 87 (29.4) 1720 (1195–2512) 2.0 (0.3–5.3)
306 (14.2; 12.7–15.6%) 66 (56–75) 205 (67.0) 33 (10.8) 38 (12.4) 174 (56.9) 93 (30.4) 271 (88.6) 41.0 (34.0–49.3) 21.5 (15.0–29.0) 8 (6–10) 150 (49.0) 222 (72.7) 137 (44.8) 2300 (1725–3749) 1.9 (−0.4–5.3)
312 (14.4; 13.0–15.9%) 68 (57–77) 207 (66.3) 45 (14.4) 54 (17.3) 184 (59.0) 117 (37.5) 273 (87.5) 51.0 (38.0–64.0) 25.0 (17.0–34.0) 10 (8–13) 167 (53.5) 262 (84.0) 141 (45.2) 787 (295–1406) 5.1 (1.8–9.7)
o0.001 0.63 o0.001 0.02 o0.001 0.01 0.03 o0.001 0.01 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001
43 72 197 172 4.5
(13.8) (23.1) (63.1) (55.1; 49.6–60.6%) (2.7–9.0)
o0.001 o0.001
129 (41.3; 35.9–46.8%)
o0.001
Highest AKI stageb Stage 1 Stage 2 Stage 3 RRT – number (%; 95% CI) Length of stay in the ICU (days)
— — — 10 (0.8; 0.3–1.3%) 2.7 (1.8–5.2)
90-day mortality – number (%; 95% CI)
235 (79.4) 47 (15.9) 14 (4.7) 20 (6.8; 3.9–9.6%) 3.8 (2.2–7.3)
207 (16.6; 14.5–18.7%)
80 (27.0; 22.0–31.1%)
157 90 59 23 3.7
(51.3) (29.4) (19.3) (7.5; 4.6–10.5%) (2.2–5.4)
60 (19.6; 15.2–24.1%)
Abbreviations: AKI, acute kidney injury; CI, confidence interval; Cr, creatinine; ICU, intensive care unit; SOFA, Sequential Organ Failure Assessment; SAPS, Simplified Acute Physiology Score; UO, urine output. a Comparison between UO-AKI, Cr-AKI, and UO and Cr AKI groups. b Cr-AKI by Cr criterion, UO-AKI by UO criterion, Cr- or UO-AKI by the one with highest stage. P-values for comparisons of separate groups. Fluid accumulation: No AKI versus Cr-AKI o0.001, No AKI versus UO-AKI o0.001, and UO-AKI versus Cr-AKI 0.85. RRT: No AKI versus Cr-AKI o0.001, No AKI versus UO-AKI o0.001, and UO-AKI versus Cr-AKI 0.75. 90-day mortality: No AKI versus Cr-AKI 0.24, No AKI versus UO-AKI o0.001, and UO-AKI versus Cr-AKI 0.03. Data presented as median (interquartile range) or number (%).
Table 2 | Incidence of separate types of oliguric episodes 3–6 h
6–12 h
12–24 h
Over 24 h
Before diagnosis of AKI by Cr/RRT or until day 5 if no AKI (n = 1966)a 0.3 to o0.5 ml/kg/h 385 (19.6%) 113 (5.7%) 6 (0.3%) 0 0.1 to o0.3 ml/kg/h 320 (16.3%) 78 (4.0%) 3 (0.2%) 0 o0.1 ml/kg/h 85 (4.3%) 38 (1.9%) 14 (0.7%) 0 During ICU stay until day 5 (n = 2160)b 0.3 to o0.5 ml/kg/h 411 (19.0%) 116 (5.4%) 0.1 to o0.3 ml/kg/h 342 (15.8%) 79 (3.7%) o0.1 ml/kg/h 110 (5.1%) 60 (2.8%)
5 (0.2%) 12 (0.6%) 58 (2.7%)
0 0 66 (3.1%)
Abbreviations: AKI, acute kidney injury; Cr, creatinine criterion; ICU, intensive care unit; RRT, renal replacement therapy. a 219 (11.1%) With no oliguria and 705 (35.9%) with oliguria lasting o3 h. b 170 (7.9%) With no oliguria and 731 (33.8%) with oliguria lasting o3 h.
Acute Physiology Score (SAPS) II score without age and renal components, use of vasoactives, diuretics, and the cumulative fluid balance (% of baseline weight) on the day of oliguric episode. The results of the regression models are presented in Table 3. UO from 0.3 to o0.5 ml/kg/h for 3–6 h was associated with a decreased risk for the development of AKI by Cr/RRT. UO o0.1 ml/kg/h for 3–6 h, 6–12 h, or 12–24 h were the only oliguric episodes that were associated with an increased risk for the development of AKI by Cr/RRT. 90-day mortality
Of the 2160 studied patients, 476 (22.0%; 95% confidence interval 20.3–23.8%) had died by day 90. Crude 90-day mortality rose along with increasing severity of oliguria Kidney International
(Figure 3). When patients treated with RRT were excluded, the 90-day mortality of patients with oliguric episodes below 12 h was comparable to the main cohort (sensitivity analysis as Supplementary Figure S1 online). In univariate analysis, oliguric episodes from 0.3 to o0.5 ml/kg/h for over 6 h and more severe episodes were associated with 90-day mortality (Table 4). We adjusted the analyses for age, sex, APACHE II diagnosis group, SAPS II score without age and renal components, use of vasoactives, diuretics, or RRT, and the cumulative fluid balance (% of baseline weight) on the day of oliguric episode. Accordingly, UO from 0.1 to o0.3 ml/kg/h for over 6 h and all episodes of UOo0.1 ml/kg/h for longer than 3 h were independently associated with an increased risk for 90-day mortality (Table 4). In addition, UO 0.3–0.5 ml/kg/h for 6–12 h (n = 116) was associated with an OR of 1.75 (95% confidence interval 1.06–2.90, P = 0.03) with an increased risk for mortality. Association of isolated oliguria with 90-day mortality
A total of 795 (36.8%) patients presented with oliguria for over 3 h but did not develop AKI by Cr/RRT. None of them had severe oliguria over 24 h. Figure 4 presents the crude 90-day mortality according to severity of oliguria and presence or absence of AKI by Cr/RRT. After multivariable adjustments among these patients with isolated oliguria compared with those with no oliguria or oliguria for o3 h, only UO from 0.1 to o0.3 ml/kg/h for over 6 h was associated with an increased risk for 90-day mortality with an 3
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
100 90 80 70 60 50
RRT Both
40
Cr-AKI
30 20 10 0
None
<3 h <0. 5ml/kg/h
219
705
3–6 h
>6 h
0.3–0.5 ml/kg/h
385
119
3–6 h
>6 h
3–6 h
0.1–0.3 ml/kg/h
320
81
6–12 h
12–24 h
<0.1 ml/kg/h
85
38
14
Figure 2 | Incidence (% with 95% confidence intervals) of acute kidney injury defined by creatinine criterion and/or renal replacement therapy according to the severity and duration of the longest oliguric episode of each patient. Numbers below each bar indicate the total number of patients in that group. Cr-AKI, acute kidney injury based on the Kidney Disease: Improving Global Outcomes creatinine criterion; RRT, renal replacement therapy.
Table 3 | Univariate and multivariatea regression models for the development of AKI based on Cr criterion or initiation of RRT No with endpoint/no with episode
Univariate odds ratio (95% CI)b
Univariate P-value
Multivariable odds ratio (95% CI)b
Multivariable P-value
UO from 0.3 to o0.5 ml/kg/h 3–6 h 59/385 (15.3%) Over 6 h 27/119 (22.7%)
0.65 (0.47–0.89) 1.05 (0.67–1.66)
0.01 0.84
0.55 (0.39–0.77) 0.38 (0.12–1.27)c
o0.001 0.12
UO from 0.1 to o0.3 ml/kg/h 3–6 h 68/320 (21.3%) Over 6 h 27/81 (33.3%)
1.07 (0.79–1.44) 1.98 (1.23–3.20)
0.66 0.01
0.84 (0.61–1.15) 1.43 (0.85–2.42)
0.28 0.18
UO o0.1 ml/kg/h 3–6 h 6–12 h 12–24 h
3.05 (1.96–4.75) 5.79 (2.99–11.20) 9.44 (2.94–30.26)
o0.001 o0.001 o0.001
2.17 (1.35–3.48) 2.95 (1.20–7.25)d 5.74 (1.67–19.73)
o0.001 0.02 0.01
38/85 (44.7%) 23/38 (60.5%) 10/14 (71.4%)
Abbreviations: AKI, acute kidney injury; RRT, renal replacement therapy; UO, urine output. a Adjusted for: age, sex, Acute Physiology And Chronic Health Evaluation II diagnosis group, Simplified Acute Physiology Score II score without age and renal components, use of vasoactives, diuretics, and the cumulative fluid balance (% of baseline weight) on the day of oliguric episode. b Results are relative to those without oliguria or with oliguria lasting o3 h. c Includes interaction term between oliguria and use of vasoactives. d Includes interaction term of presence of oliguria and cumulative fluid balance.
OR of 2.36 (95% confidence interval 1.19–4.66), P = 0.01. The median (interquartile range) duration of UO o0.3 ml/kg/h among these patients was 7.1 (6.3–8.1) h and fluid accumulation on day 1 was 1.7 (0.2–4.0)% of baseline weight.
DISCUSSION
In this large prospective multicenter study among critically ill patients we analyzed the confounder-adjusted associations of consecutive oliguria of different severity and duration with the development of AKI defined by Cr criterion and/or RRT, and secondarily, with 90-day mortality. We found that episodes of severe oliguria (UOo0.1 ml/kg/h) for more than 4
3 h were independently associated with the development of AKI by Cr/RRT. The association of severe oliguria with the development of Cr-AKI makes physiologically sense. Oliguria is defined as UO o400 ml/d (0.2–0.3 ml/kg/h), which corresponds the maximally concentrated urine volume needed to excrete the solute load normally ingested.17 Critically ill patients receive large amounts of fluid and electrolytes during the resuscitation phase, which typically leads to fluid accumulation if kidneys are unable to respond. Most patients with AKI in the current study had fluid accumulation already on day 1. Co-existing fluid retention might obscure the detection of changes in serum Cr among oliguric patients.18 Moreover, Kidney International
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
90-Day mortality
Cr-AKI/RRT
100 90 80 70 60 50 40 30 20 10 0 None
170
0.5–3 h
3–6 h
<0.5 ml/ kg/h
0.3–0.5 ml/kg/h
731
411
Over 6 h
121
3–6 h
Over 6 h
3–6 h
0.1–0.3 ml/kg/h 342
91
6–12 h
12–24 h Over 24 h
<0.1 ml/kg/h 110
60
58
66
Figure 3 | 90-day mortality and presence of acute kidney injury defined by creatinine criterion and/or use of renal replacement therapy until day 5 in the ICU (% with 95% confidence intervals) according to the severity and duration of the longest oliguric episode of each patient. Numbers below each bar indicate the total number of patients in that group. The UO data of patients with RRT were screened also during RRT. Cr-AKI, acute kidney injury based on the Kidney Disease: Improving Global Outcomes creatinine criterion; RRT, renal replacement therapy; UO, urine output.
Table 4 | Univariate and multivariablea regression models for 90-day mortality Crude 90-day mortalityb
Univariate odds ratio (95% CI)c
Univariate P-value
Multivariable odds ratio (95% CI)c
Multivariable P-value
UO from 0.3 to o0.5 ml/kg/h 3–6 h 69/411 (16.8%) Over 6 h 31/121 (25.6%)
1.06 (0.078–1.45) 1.81 (1.16–2.83)
0.71 0.01
0.96 (0.68–1.37) 1.65 (1.00–2.72)
0.83 0.05
UO from 0.1 to o0.3 ml/kg/h 3–6 h 78/342 (22.8%) Over 6 h 26/91 (28.6%)
1.44 (1.08–1.92) 1.95 (1.21–3.14)
0.01 0.01
1.20 (0.87–1.65) 1.96 (1.13–3.38)
0.28 0.02
UO o0.1 ml/kg/h 3–6 h 34/110 (30.9%) 6–12 h 24/60 (40.0%) 12–24 h 31/58 (53.4%) Over 24 h 39/66 (59.1%)
1.95 2.91 5.01 6.30
(1.28–2.97) (1.71–4.94) (2.95–8.50) (3.81–10.43)
0.01 o0.001 o0.001 o0.001
2.08 3.04 6.78 9.76
(1.27–3.42) (1.50–6.15) (2.79–16.44) (4.07–23.43)
0.01 0.01 o0.001 o0.001
Abbreviations: RRT, renal replacement therapy; UO, urine output. a Adjusted for: age, sex, Acute Physiology And Chronic Health Evaluation II diagnosis group, Simplified Acute Physiology Score II score without age and renal components, use of vasoactives, diuretics, and RRT, and the cumulative fluid balance (% of baseline weight) on the day of oliguric episode. b Expressed as number/total number of patients with episode (%). c Results are relative to those without oliguria or with oliguria lasting o3 h.
because ICU patients generally do not receive full nutrition during their first days in the ICU, the amount of Cr to be excreted may be lower. These factors may explain why only severe oliguria as a sign of very low GFR was associated with AKI by Cr/RRT. We found the results regarding isolated oliguria interesting. First, patients with oliguric AKI but no Cr-AKI had Kidney International
higher crude mortality compared with those without any AKI. These results corroborate those of a database analysis reporting isolated stage 2–3 UO AKI to be associated with decreased age-adjusted one-year survival.11 Second, moderate isolated oliguria for over 6 h was independently associated with 90-day mortality. Notably, the median duration of oliguria among these patients was about 7 h. Patients 5
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
With Cr-AKI/RRT
No Cr-AKI/RRT
80 70 60 50 40 30 20 10 0 None
0.5–3 h <0.5 ml/kg/ h
30 140
147 584
3–6 h
Over 6 h
0.3–0.5 ml/kg/h 82 329
29 92
3–6 h
Over 6 h
3–6 h
0.1–0.3 ml/kg/h 90 252
37 54
6–12 h
12–24 h
Over 24 h
<0.1 ml/kg/h 61 49
45 15
54 4
66 0
Figure 4 | Comparison of 90-day mortality (% with 95% confidence intervals) of patients who developed acute kidney injury diagnosed by creatinine-criterion and/or use of renal replacement therapy with those who did not. Data are presented according to the severity and duration of the longest oliguric episode of each patient. Numbers below each bar indicate the total number of patients in that group. Cr-AKI, acute kidney injury based on the Kidney Disease: Improving Global Outcomes creatinine criterion; RRT, renal replacement therapy.
presenting with severe isolated oliguria were few and the sample size may have been too small to show a significant association among them. Again, fluid accumulation may explain why these patients did not develop Cr-AKI. However, moderate isolated oliguria might be a sign of primarily non-renal derangements (e.g. hemorrhagic shock) that if resolved, may not ever lead to a rise in Cr or prolonged oliguria. The primary cause of these derangements may still expose the patients to adverse outcome. Thus, oliguria is an important warning sign regardless of later development of Cr-AKI. In our analysis, very few patients with severe oliguria did not develop Cr-AKI, but only a third of patients with mild or moderate oliguria developed Cr-AKI or received RRT. Notably, patients were diagnosed with AKI by Cr/RRT very early, after a median of 16 h in the ICU. Previously almost 50% of episodes of UO o0.5 ml/kg/h for over 4 h have been reported not to be followed by Cr-AKI.14 Although the incidence of Cr-AKI/RRT rose with increasing severity of oliguria, over a third of patients without oliguria developed Cr-AKI. Potentially, the etiology and clinical course for AKI may be different in Cr-AKI only patients compared with those presenting with oliguria also. Surprisingly, patients with Cr-AKI only did not have significantly higher crude 90-day mortality compared with patients without AKI of any type. This is in contrast with the findings by Kellum et al.11 However, the 90-day mortality of non-AKI patients in our analysis was higher (16.6%) compared with 7.3% reported previously,11 which is plausibly due to exclusion of patients 6
with scheduled post-surgery admission o24 h from the FINNAKI study. We defined the duration of oliguria as consecutive hours below the studied threshold according to the strictest interpretation of the UO criterion.4,5 The distinct ways of defining oliguria may explain the controversy of our results with a previous a study that did not find UO o0.5 ml/kg/h for 6 h to be associated with mortality after adjustments.16 In their single-center study comprising 725 patients they calculated UO in moving blocks during the first 24 h in the ICU and used the 6-hour period with lowest UO to define and stage AKI.16 Averaging the UO over a 6-hour collection period has been shown to be more sensitive than entailing consecutive oliguric hours but reduce the specificity of UO in predicting the development of AKI based on the AKI Network (AKIN) Cr criterion.8 To the best of our knowledge, this is the first prospective multicenter study to evaluate the association of hourly measured UO with a fixed mortality end-point. We were able to separate three different degrees of consecutive oliguria and study their independent association, first with development of AKI by Cr/, and second with 90-day mortality. Even mild oliguria for 6–12 h was independently associated with an increased risk for mortality, thus providing support for the adequacy of the current definition of oliguria. Moreover, our results corroborate those reported by Kellum et al.11 regarding that patients with both Cr-AKI and UO-AKI had the highest rates of RRT and 90-day mortality and that also isolated oliguria was associated with an adverse outcome. These Kidney International
ST Vaara et al.: Oliguria in the critically ill
findings underlie the importance of hourly UO measurements and necessity of using UO criterion along with Cr criterion to diagnose AKI. The strengths of our study include the prospective multicenter design, multiple data on confounding factors, and a fixed-time mortality. Several limitations, however, should be addressed. First, we did not record the volume status of oliguric patients at the start of the oliguric episode, and possibly, some of the patients may have been dehydrated. However, we adjusted for both daily fluid balance and use of diuretics in the multivariable models. Second, unmeasured confounders cannot be excluded despite that we adjusted for a number of important confounders. Third, we could not discern the potential effects of initiation of RRT on the UO but handled the data of RRT and non-RRT patients similarly regarding the mortality outcome. Fourth, as in previous studies on this topic,8,14,16 we excluded patients with an ICU length of stay shorter than 24 h to enable studying also longer periods of oliguria. Therefore, the overall mortality rate in this analysis was lower than in the main FINNAKI study.19 Fifth, among 80% of the patients, the baseline weight was reported by the patient or his/hers proxy or estimated by experienced personnel. We used the weight recorded at ICU admission as the baseline weight. No data which weight recording (baseline, actual, or ideal) would be the most accurate in assessing urine flow rate exist.6 Because most of the patients presented with their longest episode of oliguria on day 1 when the fluid accumulation was moderate in all patient groups (Table 1), we believe that using baseline weight instead of actual daily measured weight has not caused significant bias. Had we used the actual weight, the incidence of oliguria could be slightly greater. Baseline weight along with partly assessed patient weights were used also previously.16 Finally, the sample size of this study did not allow repeating the analyses in the subcohort of patients admitted with severe sepsis or in patients without diuretics. Despite these limitations, we consider the external validity of this study to be good among critically ill patients of Caucasian origin admitted to general ICUs. In conclusion, we demonstrated that episodes of severe oliguria (UO o0.1 ml/kg/h) for more than 3 h were independently associated with the development of Cr-AKI and/or administration of RRT. Patients with isolated oliguric AKI had higher crude mortality compared with patients without AKI. Our findings underlie the importance of hourly measurement of diuresis. MATERIALS AND METHODS This study was a secondary analysis using the prospectively collected data from the FINNAKI Study conducted in 17 Finnish ICUs between 1 September 2011 and 1 February 2012.19 The Ethics committee of Department of Surgery, Helsinki and Uusimaa Hospital District approved the study protocol and the use of deferred consent with written informed consent obtained from the patient or proxy as soon as possible. The Finnish National Kidney International
clinical investigation
Institute of Health approved collection of data from medical records of deceased patients if informed consent could not be obtained. The study was conducted according to the declaration of Helsinki. Patients All patients with either an emergency admission to ICU of any duration or post-surgical scheduled admission expected to exceed 24 h were eligible in the FINNAKI study if none of the following exclusion criteria were met: (1) pre-existing end-stage renal disease and need for maintenance dialysis, (2) previous receipt of RRT while enrolled in the study, (3) organ donor, (4) intermediate care patient, (5) patient not permanently living in Finland, (6) patients transferred from another participating ICU after full study data collection period, and (7) absence of consent. From the current analysis, we further excluded several patient groups presented in the study flow chart (Figure 1). Data collection We recorded data from patients’ pre-existing conditions, medications, and ICU treatment with case report forms and obtained physiological data including UO, laboratory values, diagnoses, and severity of illness scores from the ICU data management system via the Finnish Intensive Care Consortium platform. The data collection period was until day 5 in the ICU. All patients had urine catheters and UO was registered hourly in the electronic patient charts. Patient weight recording at ICU admission in all participating centers was based either on weighting (in the ICU or preoperatively in 426 patients (19.7%)) or assessed (according to the reported weight by the patient or his/her proxy or estimated bedside by personnel) in 1734 (80.3%). The baseline weight determined at ICU admission was used throughout the study to calculate urinary flow rates. The intensivists screened patients daily for the presence of severe sepsis according to the ACCP/SCCM definition.20 We used the KDIGO criteria to screen for and stage AKI.6 Plasma Cr measurements were performed in accredited hospital laboratories using enzymatic method with an interassay coefficient of variation o3–4%. As baseline Cr, we used the latest value obtained within 1 year but at least a week before study inclusion. If baseline Cr value could not be obtained, we estimated it using the Modification of Diet in Renal Disease equation assuming a GFR of 75 ml/min/1.73 m2 as recommended.4 Oliguria An electronic calculator was constructed to calculate the UO ml/kg/h between two UO recordings. We screened these data for the presence of the following degrees of severity oliguria: (1) o0.5 ml/kg/h, (2) o0.3 ml/kg/h, and (3) o0.1 ml/kg/h. Regarding the development of AKI by Cr/RRT, we screened the data before fulfillment of this endpoint. Regarding 90-day mortality, we screened data until ICU discharge or day 5 in the ICU, and included UO data obtained during RRT among those with RRT. We recorded the duration of each episode of consecutive values below the threshold. We formed exclusive categories of the severity of oliguria (mild oliguria from 0.3 to o0.5 ml/kg/h, moderate oliguria from 0.1 to o0.3 ml/kg/h) by classifying the patients according to their longest oliguric episode without periods of deeper oliguria more than 3 h. Definition of severe oliguria was UO o0.1 ml/kg/h. The duration of episodes were categorized into o3 h, 3–6 h, 6–12 h, 12–24 h, and longer than 24 h. 7
clinical investigation
Association of oliguria with the endpoints We used logistic regression to study the association between the different types of oliguric episodes with the development of AKI by Cr/RRT, and secondarily with 90-day mortality. We adjusted for the following a priori selected confounders: (1) fluid accumulation (%) of baseline weight, (2) use of diuretics, (3) use of vasoactive medication, (4) SAPS II score (day 1) without points given for age and renal components, (5) age, (6) sex, and (7) APACHE II admission diagnosis group. For confounders 1–3, we used values obtained on the day of the beginning of the oliguric episode. Regarding 90-day mortality, we also adjusted for the use of RRT. We calculated the median day number of occurrence of each type of oliguric episode, and used data on that day for adjustments among patients without the tested type of oliguric episode. We used multiplicative terms to test for presence of interactions between confounders 1 and 3 and use of RRT with oliguria. We assessed the model fit with the Hosmer–Lemeshow test.
ST Vaara et al.: Oliguria in the critically ill
6.
7. 8. 9.
10.
11. 12.
13.
14.
Statistical analysis We report categorical variables with count and percentage and continuous variables with median and interquartile range. We used Fischer’s exact test to compare categorical variables and Mann–Whitney U-test to compare continuous variables, or in case of comparing groups of three or more χ2 test or Kruskal–Wallis test. We used logistic regression with enter method to adjust for confounders. We set the significance level at 0.05 and report two-tailed P-values. We conducted the analysis with SPSS statistics 20.0. (IBM, Armonk, NY, USA). DISCLOSURE
All the authors declared no competing interests. ACKNOWLEDGMENTS
We are grateful to professor Seppo Sarna for valuable statistical advice. We thank Tieto Healthcare & Welfare for database management. The Academy of Finland, the Sigrid Juselius Foundation, Päivikki and Sakari Sohlberg Foundation, and Helsinki University Central Hospital EVO-grants have supported this study. SUPPLEMENTARY MATERIAL Figure S1. The 90-day mortality (%, 95% confidence interval) according to the longest oliguric episode of patients who did not require RRT. Supplementary material is linked to the online version of the paper at http://www.nature.com/ki. REFERENCES 1.
2.
3.
4.
5.
8
Uchino S, Kellum JA, Bellomo R et al. Acute renal failure in critically ill patients: a multinational, multicenter study. JAMA 2005; 294: 813–818. Chertow GM, Burdick E, Honour M et al. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J Am Soc Nephrol 2005; 16: 3365–3370. Joannidis M, Metnitz B, Bauer P et al. Acute kidney injury in critically ill patients classified by AKIN versus RIFLE using the SAPS 3 database. Intensive Care Med 2009; 35: 1692–1702. Bellomo R, Ronco C, Kellum JA et al. Acute renal failure - definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004; 8: R204–R212. Mehta RL, Kellum JA, Shah SV et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007; 11: R31.
15.
16. 17. 18.
19.
20.
Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group: KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl 2012; 2: 1–138. Ricci Z, Cruz D, Ronco C. The RIFLE criteria and mortality in acute kidney injury: a systematic review. Kidney Int 2008; 73: 538–546. Macedo E, Malhotra R, Bouchard J et al. Oliguria is an early predictor of higher mortality in critically ill patients. Kidney Int 2011; 80: 760–767. McIlroy DR, Argenziano M, Farkas D et al. Incorporating oliguria into the diagnostic criteria for acute kidney injury after on-pump cardiac surgery: impact on incidence and outcomes. J Cardiothorac Vasc Anesth 2013; 27: 1145–1152. Wlodzimirow KA, Abu-Hanna A, Slabbekoorn M et al. A comparison of RIFLE with and without urine output criteria for acute kidney injury in critically ill patients. Crit Care 2012; 16: R200. Kellum JA, Sileanu FE, Murugan R et al. Classifying AKI by Urine Output versus SerumCreatinine Level. J Am Soc Nephrol 2015. Mandelbaum T, Lee J, Scott DJ et al. Empirical relationships among oliguria, creatinine, mortality, and renal replacement therapy in the critically ill. Intensive Care Med 2013; 39: 414–419. Mandelbaum T, Scott DJ, Lee J et al. Outcome of critically ill patients with acute kidney injury using the Acute Kidney Injury Network criteria. Crit Care Med 2011; 39: 2659–2664. Prowle JR, Liu YL, Licari E et al. Oliguria as predictive biomarker of acute kidney injury in critically ill patients. Crit Care 2011; 15: R172. Leedahl DD, Frazee EN, Schramm GE et al. Derivation of urine output thresholds that identify a very high risk of AKI in patients with septic shock. Clin J Am Soc Nephrol 2014; 9: 1168–1174. Md Ralib A, Pickering JW, Shaw GM et al. The urine output definition of acute kidney injury is too liberal. Crit Care 2013; 17: R112. Klahr S, Miller SB. Acute oliguria. N Engl J Med 1998; 338: 671–675. Liu KD, Thompson BT, Ancukiewicz M et al. Acute kidney injury in patients with acute lung injury: impact of fluid accumulation on classification of acute kidney injury and associated outcomes. Crit Care Med 2011; 39: 2665–2671. Nisula S, Kaukonen KM, Vaara ST et al. Incidence, risk factors and 90-day mortality of patients with acute kidney injury in Finnish intensive care units: the FINNAKI study. Intensive Care Med 2013; 39: 420–428. Bone RC, Balk RA, Cerra FB et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/ SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992; 101: 1644–1655.
FINNAKI STUDY GROUP
Central Finland Central Hospital: Raili Laru-Sompa, Anni Pulkkinen, Minna Saarelainen, Mikko Reilama, Sinikka Tolmunen, Ulla Rantalainen, Marja Miettinen East Savo Central Hospital: Markku Suvela, Katrine Pesola, Pekka Saastamoinen, Sirpa Kauppinen Helsinki University Central Hospital: Ville Pettilä, Kirsi-Maija Kaukonen, Anna-Maija Korhonen, Sara Nisula, Suvi Vaara, Raili Suojaranta-Ylinen, Leena Mildh, Mikko Haapio, Laura Nurminen, Sari Sutinen, Leena Pettilä, Helinä Laitinen, Heidi Syrjä, Kirsi Henttonen, Elina Lappi, Hillevi Boman Jorvi Central Hospital: Tero Varpula, Päivi Porkka, Mirka Sivula, Mira Rahkonen, Anne Tsurkka, Taina Nieminen, Niina Pirttinen Kanta-Häme Central hospital: Ari Alaspää, Ville Salanto, Hanna Juntunen, Teija Sanisalo Kuopio University Hospital: Ilkka Parviainen, Ari Uusaro, Esko Ruokonen, Stepani Bendel, Niina Rissanen, Maarit Lång, Sari Rahikainen, Saija Rissanen, Merja Ahonen, Elina Halonen, Eija Vaskelainen Lapland Central Hospital: Meri Poukkanen, Esa Lintula, Sirpa Suominen Länsi-Pohja Central Hospital: Jorma Heikkinen, Timo Lavander, Kirsi Heinonen, Anne-Mari Juopperi Middle Ostrobothnia Central Hospital: Tadeusz Kaminski, Fiia Gäddnäs, Tuija Kuusela, Jane Roiko North Karelia Central Hospital: Sari Karlsson, Matti Reinikainen, Tero Surakka, Helena Jyrkönen, Tanja Kidney International
ST Vaara et al.: Oliguria in the critically ill
Eiserbeck, Jaana Kallinen Oulu University Hospital: Tero AlaKokko, Jouko Laurila, Sinikka Sälkiö Satakunta Hospital District: Vesa Lund, Päivi Tuominen, Pauliina Perkola, Riikka Tuominen, Marika Hietaranta, Satu Johansson South Karelia Central Hospital: Seppo Hovilehto, Anne Kirsi, Pekka Tiainen, Tuija Myllärinen, Pirjo Leino, Anne Toropainen
Kidney International
clinical investigation
Tampere University Hospital: Anne Kuitunen, Jyrki Tenhunen, Ilona Leppänen, Markus Levoranta, Sanna Hoppu, Jukka Sauranen, Atte Kukkurainen, Samuli Kortelainen, Simo Varila Turku University Hospital: Outi Inkinen, Niina Koivuviita, Jutta Kotamäki, Anu Laine Vaasa Central Hospital: SimoPekka Koivisto, Raku Hautamäki, Maria Skinnar.
9
clinical investigation
© 2015 International Society of Nephrology
Association of oliguria with the development of acute kidney injury in the critically ill Suvi T. Vaara1, Ilkka Parviainen2, Ville Pettilä1, Sara Nisula1, Outi Inkinen3, Ari Uusaro2 and The FINNAKI Study Group4 1
Intensive Care Medicine, Department of Perioperative, Intensive Care and Pain Medicine, Helsinki University Central Hospital, Helsinki, Finland; 2Division of Intensive Care, Kuopio University Hospital, Kuopio, Finland and 3Department of Intensive Care, Turku University Hospital, Turku, Finland
Urine output (UO) criterion may increase the sensitivity of the definition of acute kidney injury (AKI). We determined whether the empirically derived definition for oliguria (o0.5 ml/kg/h) is independently associated with adverse outcome. Data analysis included hourly recorded UO from the prospective, multicenter FINNAKI study conducted in 16 Finnish intensive care units. Confounder-adjusted association of oliguria of different severity and duration primarily with the development of AKI defined by creatinine criterion (Cr-AKI) or renal replacement therapy (RRT) was assessed. Secondarily, we determined the association of oliguria with 90-day mortality. Of the 1966 patients analyzed for the development of AKI, 454 (23.1%) reached this endpoint. Within this AKI cohort, 312 (68.7%) developed Cr-AKI, 21 (4.6%) commenced RRT without Cr-AKI, and 121 (26.7%) commenced RRT with Cr-AKI. Episodes of severe oliguria (o0.1 ml/kg/h) for more than 3 h were independently associated with the development of Cr-AKI or RRT. The shortest periods of consecutive oliguria independently associated with an increased risk for 90-day mortality were 6–12 h of oliguria from 0.3 to o0.5 ml/kg/h, over 6 h of oliguria from 0.1 to o0.3 ml/kg/h, and severe oliguria lasting over 3 h. Thus, our findings underlie the importance of hourly UO measurements. Kidney International advance online publication, 9 September 2015; doi:10.1038/ki.2015.269
Correspondence: Suvi T. Vaara, Intensive Care Medicine, Department of Perioperative, Intensive Care and Pain Medicine, Helsinki University Central Hospital, Box 340, 00290 Helsinki, Finland. E-mail: [email protected] 4
See appendix.
Received 27 October 2014; revised 3 July 2015; accepted 9 July 2015 Kidney International
Acute kidney injury (AKI) is a frequently diagnosed syndrome in the intensive care unit (ICU) that associates with increased mortality.1–3 The diagnosis of AKI is based on an abrupt increase in plasma creatinine (Cr) and/or a decrease in urine output (UO).4–6 Even small increases in Cr are associated with increased short- and long-term mortality and morbidity.2,6 Cr criteria for AKI have been extensively studied7 and recently updated.6 On contrary, the UO criterion has been empirically derived4,5 and its further validation has been suggested.5 Few studies, however, have been able to employ it without modifications. Combining UO and Cr criteria increases the sensitivity of the AKI definition.3,8–10 A recent database analysis showed that patients with both UO and Cr criteria had highest short- and long-term mortality rates.11 Isolated oliguria without fulfillment of Cr criterion also associated with long-term mortality11 and in some studies UO has outperformed Cr in mortality prediction.12,13 Oliguria (UO o0.5 ml/kg/h) has been shown to associate with occurrence of AKI by Risk, Injury, Failure, Loss of Function, and End-stage Renal Disease Cr criterion, but to be only a fair predictor of consequent AKI.14 Among patients with septic shock, consecutive oliguria of 3–5 h has been demonstrated to predict occurrence of Kidney Disease: Improving Global Outcomes (KDIGO) stage 2 or 3 AKI by Cr criterion.15 However, a single center study found that UO o0.3 ml/kg/h for 6 h corresponded better the current Cr-based definition of AKI than UO o0.5 ml/kg/h for 6 h, and was associated with mortality after adjustments whereas UO o0.5 ml/kg/h was not.16 Moreover, UO is prone to confounders such as use of diuretics, differences in fluid management and balance, and in administration of renal replacement therapy (RRT). In this large prospective multicenter study we aimed to find out whether the empirically derived threshold for oliguria o0.5 ml/kg/h is independently associated with adverse outcome. We studied the confounder-adjusted association of oliguria of different severity (mild oliguria from 0.3 to o0.5 ml/kg/h, moderate oliguria from 0.1 to o0.3 ml/kg/h, and severe oliguria o0.1 ml/kg/h) and their duration (3–6 h, 6–12 h, 12–24 h, and over 24 h) 1
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
2901 Patients in the FINNAKI study Exclusions: - 560 With incomplete UO recording within first 24 h - 6 Without baseline weight - 11 With renal transplant - 9 With acute RRT already before ICU admission - 96 With UO recordings done using different method - 63 With incomplete data - 188 Readmissions
2160 Patients analyzed regarding 90-day mortality -90-day mortality 476 (22.0%; 95% CI 20.3–23.8%) 194 (8.9%) Excluded with Cr-AKI/ RRT within 3 h of ICU admission
1966 Patients analyzed regarding Cr-AKI/RRT endpoint -endpoint positive in 454 (23.1%; 95% CI 21.2–25.0%)
Figure 1 | Study flow-chart with excluded patient groups. CI, confidence interval; Cr-AKI, acute kidney injury by creatinine criterion; ICU, intensive care unit; RRT, renal replacement therapy; UO, urine output.
primarily with the development of AKI defined by Cr criterion (Cr-AKI) or administration of RRT. Secondarily, we studied the association of oliguria with 90-day mortality. RESULTS
After exclusions presented in the study flow chart (Figure 1), 2160 patients were included in the analysis. Altogether 924 patients (42.8%) fulfilled the KDIGO criteria (UO, Cr, or RRT) for AKI. Of all, 225 (10.4%) patients received RRT, and 222 patients commenced RRT within the first 24 h in the ICU. The 90-day mortality (95% confidence interval) of all AKI patients was 29.3% (26.4–32.3%). Patients, who fulfilled both Cr and UO criteria, were most severely ill and had the highest rates for administration of RRT and 90-day mortality (Table 1). Oliguria
At least one episode of oliguria (UO below 0.5 ml/kg/h for a minimum of 0.5 h) occurred in 1990 (92.1%) patients. Episodes longer than 30 min of moderate oliguria (UO o0.3 ml/kg/h) occurred in 1694 (78.4%) of the patients and of severe oliguria (UO o0.1 ml/kg/h) in 900 (41.7%) of patients. The median (interquartile range) day of the beginning of the longest oliguric episode was day 1 (0–2). Altogether 608 (28.1%) patients had oliguria with UO o0.5 ml/kg/h for ⩾ 6 consecutive hours, and thus, oliguric AKI. Of these, 330 (54.3%) had stage 1 AKI, 132 (21.7%) stage 2 AKI, and 146 (24.0%) stage 3 AKI according to the KDIGO UO criterion. As short episodes of oliguria of all severities were very frequent, categories of UO from 0.3 to o0.5 ml/kg/h and from 0.1 to o0.3 ml/kg/h were formed by excluding patients with more severe oliguria for longer than 3 h. The incidence of oliguric episodes (the longest episode of each patient included) of different degrees of severity is presented in Table 2. 2
Development of AKI defined by Cr criterion and/or administration of RRT
We studied the association of oliguria and development of AKI based on Cr criterion and/or initiation of RRT (AKI by Cr/RRT) in 1966 patients. Because 194 patients developed AKI by Cr/RRT within 3 h from ICU admission, they were excluded from the analysis regarding the development of AKI. UO data were screened until the diagnosis of AKI by Cr/RRT and among those who did not develop AKI, until ICU discharge or day 5. Of the 1966 patients, 454 (23.1%) developed AKI by Cr/RRT, 312 (68.7%) of them developed Cr-AKI, 21 (4.6%) commenced RRT without Cr-AKI, and 121 (26.7%) commenced RRT with Cr-AKI. The median (interquartile range) time to reach AKI diagnosis by Cr/RRT was 16.9 (10.7–34.3) hours. Altogether 1747 (88.9%) patients had at least one oliguric episode lasting over 30 min. Oliguric AKI was diagnosed in 147 (32.4%) patients before the diagnosis of AKI by Cr/RRT and in 276 (18.3%) of those who did not develop AKI by Cr/RRT criteria (Po0.001). Episodes of UO o0.5 ml/kg/h for more than 12 h occurred in 94 (4.8%) patients, and 52 (55.3%) of them developed AKI by Cr/RRT. None of the patients presented with oliguria lasting over 24 h. Figure 2 presents the proportions of patients who developed AKI by Cr/RRT according to the severity and duration of oliguria. Of the 1042 patients with UO o0.5 ml/ kg/h for more than 3 h, only 248 (23.8%) did not receive diuretics in the ICU. We constructed logistic regression models to study the independent association of different types of oliguric episodes with the development of AKI by Cr/RRT. The reference group comprised those with no oliguria or oliguric episodes lasting o3 h. We adjusted for age, sex, Acute Physiology and Chronic Health Evaluation (APACHE) II diagnosis group, Simplified Kidney International
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
Table 1 | Incidence, patient characteristics, and outcomes according to the presence of UO and/or Cr criteria for AKI No of patients – number (%; 95% CI) Age Male sex Chronic kidney disease Chronic heart failure Hypertension Operative admission Emergency admission SAPS II score Non-renal non-age SAPS II score SOFA score, day 1 Severe sepsis diagnosed by day 5 Use of vasoactives (admission day) Use of diuretics (admission day) Cumulative UO within first 24 h in ICU (ml/kg/h) Fluid accumulation % of baseline weight on day 1
No AKI
UO-AKI
Cr-AKI
Cr and UO AKI
P-valuea
1246 (57.7; 55.6–59.8%) 63 (52–73) 795 (63.8) 43 (3.5) 138 (11.1) 548 (44.0) 495 (39.7) 1036 (83.1) 34.0 (27.0–44.0) 21.0 (14.0–30.0) 7 (4–8) 317 (25.4) 732 (58.7) 357 (28.7) 2500 (1859–3546) 1.4 (−1.1–3.7)
296 (13.7; 12.3–15.2%) 64 (54–73) 196 (66.2) 21 (7.1) 35 (11.8) 146 (49.3) 123 (41.6) 247 (83.4) 36.5 (29.0–50.0) 23.0 (16.0–33.0) 7 (6–9) 82 (27.7) 221 (74.7) 87 (29.4) 1720 (1195–2512) 2.0 (0.3–5.3)
306 (14.2; 12.7–15.6%) 66 (56–75) 205 (67.0) 33 (10.8) 38 (12.4) 174 (56.9) 93 (30.4) 271 (88.6) 41.0 (34.0–49.3) 21.5 (15.0–29.0) 8 (6–10) 150 (49.0) 222 (72.7) 137 (44.8) 2300 (1725–3749) 1.9 (−0.4–5.3)
312 (14.4; 13.0–15.9%) 68 (57–77) 207 (66.3) 45 (14.4) 54 (17.3) 184 (59.0) 117 (37.5) 273 (87.5) 51.0 (38.0–64.0) 25.0 (17.0–34.0) 10 (8–13) 167 (53.5) 262 (84.0) 141 (45.2) 787 (295–1406) 5.1 (1.8–9.7)
o0.001 0.63 o0.001 0.02 o0.001 0.01 0.03 o0.001 0.01 o0.001 o0.001 o0.001 o0.001 o0.001 o0.001
43 72 197 172 4.5
(13.8) (23.1) (63.1) (55.1; 49.6–60.6%) (2.7–9.0)
o0.001 o0.001
129 (41.3; 35.9–46.8%)
o0.001
Highest AKI stageb Stage 1 Stage 2 Stage 3 RRT – number (%; 95% CI) Length of stay in the ICU (days)
— — — 10 (0.8; 0.3–1.3%) 2.7 (1.8–5.2)
90-day mortality – number (%; 95% CI)
235 (79.4) 47 (15.9) 14 (4.7) 20 (6.8; 3.9–9.6%) 3.8 (2.2–7.3)
207 (16.6; 14.5–18.7%)
80 (27.0; 22.0–31.1%)
157 90 59 23 3.7
(51.3) (29.4) (19.3) (7.5; 4.6–10.5%) (2.2–5.4)
60 (19.6; 15.2–24.1%)
Abbreviations: AKI, acute kidney injury; CI, confidence interval; Cr, creatinine; ICU, intensive care unit; SOFA, Sequential Organ Failure Assessment; SAPS, Simplified Acute Physiology Score; UO, urine output. a Comparison between UO-AKI, Cr-AKI, and UO and Cr AKI groups. b Cr-AKI by Cr criterion, UO-AKI by UO criterion, Cr- or UO-AKI by the one with highest stage. P-values for comparisons of separate groups. Fluid accumulation: No AKI versus Cr-AKI o0.001, No AKI versus UO-AKI o0.001, and UO-AKI versus Cr-AKI 0.85. RRT: No AKI versus Cr-AKI o0.001, No AKI versus UO-AKI o0.001, and UO-AKI versus Cr-AKI 0.75. 90-day mortality: No AKI versus Cr-AKI 0.24, No AKI versus UO-AKI o0.001, and UO-AKI versus Cr-AKI 0.03. Data presented as median (interquartile range) or number (%).
Table 2 | Incidence of separate types of oliguric episodes 3–6 h
6–12 h
12–24 h
Over 24 h
Before diagnosis of AKI by Cr/RRT or until day 5 if no AKI (n = 1966)a 0.3 to o0.5 ml/kg/h 385 (19.6%) 113 (5.7%) 6 (0.3%) 0 0.1 to o0.3 ml/kg/h 320 (16.3%) 78 (4.0%) 3 (0.2%) 0 o0.1 ml/kg/h 85 (4.3%) 38 (1.9%) 14 (0.7%) 0 During ICU stay until day 5 (n = 2160)b 0.3 to o0.5 ml/kg/h 411 (19.0%) 116 (5.4%) 0.1 to o0.3 ml/kg/h 342 (15.8%) 79 (3.7%) o0.1 ml/kg/h 110 (5.1%) 60 (2.8%)
5 (0.2%) 12 (0.6%) 58 (2.7%)
0 0 66 (3.1%)
Abbreviations: AKI, acute kidney injury; Cr, creatinine criterion; ICU, intensive care unit; RRT, renal replacement therapy. a 219 (11.1%) With no oliguria and 705 (35.9%) with oliguria lasting o3 h. b 170 (7.9%) With no oliguria and 731 (33.8%) with oliguria lasting o3 h.
Acute Physiology Score (SAPS) II score without age and renal components, use of vasoactives, diuretics, and the cumulative fluid balance (% of baseline weight) on the day of oliguric episode. The results of the regression models are presented in Table 3. UO from 0.3 to o0.5 ml/kg/h for 3–6 h was associated with a decreased risk for the development of AKI by Cr/RRT. UO o0.1 ml/kg/h for 3–6 h, 6–12 h, or 12–24 h were the only oliguric episodes that were associated with an increased risk for the development of AKI by Cr/RRT. 90-day mortality
Of the 2160 studied patients, 476 (22.0%; 95% confidence interval 20.3–23.8%) had died by day 90. Crude 90-day mortality rose along with increasing severity of oliguria Kidney International
(Figure 3). When patients treated with RRT were excluded, the 90-day mortality of patients with oliguric episodes below 12 h was comparable to the main cohort (sensitivity analysis as Supplementary Figure S1 online). In univariate analysis, oliguric episodes from 0.3 to o0.5 ml/kg/h for over 6 h and more severe episodes were associated with 90-day mortality (Table 4). We adjusted the analyses for age, sex, APACHE II diagnosis group, SAPS II score without age and renal components, use of vasoactives, diuretics, or RRT, and the cumulative fluid balance (% of baseline weight) on the day of oliguric episode. Accordingly, UO from 0.1 to o0.3 ml/kg/h for over 6 h and all episodes of UOo0.1 ml/kg/h for longer than 3 h were independently associated with an increased risk for 90-day mortality (Table 4). In addition, UO 0.3–0.5 ml/kg/h for 6–12 h (n = 116) was associated with an OR of 1.75 (95% confidence interval 1.06–2.90, P = 0.03) with an increased risk for mortality. Association of isolated oliguria with 90-day mortality
A total of 795 (36.8%) patients presented with oliguria for over 3 h but did not develop AKI by Cr/RRT. None of them had severe oliguria over 24 h. Figure 4 presents the crude 90-day mortality according to severity of oliguria and presence or absence of AKI by Cr/RRT. After multivariable adjustments among these patients with isolated oliguria compared with those with no oliguria or oliguria for o3 h, only UO from 0.1 to o0.3 ml/kg/h for over 6 h was associated with an increased risk for 90-day mortality with an 3
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
100 90 80 70 60 50
RRT Both
40
Cr-AKI
30 20 10 0
None
<3 h <0. 5ml/kg/h
219
705
3–6 h
>6 h
0.3–0.5 ml/kg/h
385
119
3–6 h
>6 h
3–6 h
0.1–0.3 ml/kg/h
320
81
6–12 h
12–24 h
<0.1 ml/kg/h
85
38
14
Figure 2 | Incidence (% with 95% confidence intervals) of acute kidney injury defined by creatinine criterion and/or renal replacement therapy according to the severity and duration of the longest oliguric episode of each patient. Numbers below each bar indicate the total number of patients in that group. Cr-AKI, acute kidney injury based on the Kidney Disease: Improving Global Outcomes creatinine criterion; RRT, renal replacement therapy.
Table 3 | Univariate and multivariatea regression models for the development of AKI based on Cr criterion or initiation of RRT No with endpoint/no with episode
Univariate odds ratio (95% CI)b
Univariate P-value
Multivariable odds ratio (95% CI)b
Multivariable P-value
UO from 0.3 to o0.5 ml/kg/h 3–6 h 59/385 (15.3%) Over 6 h 27/119 (22.7%)
0.65 (0.47–0.89) 1.05 (0.67–1.66)
0.01 0.84
0.55 (0.39–0.77) 0.38 (0.12–1.27)c
o0.001 0.12
UO from 0.1 to o0.3 ml/kg/h 3–6 h 68/320 (21.3%) Over 6 h 27/81 (33.3%)
1.07 (0.79–1.44) 1.98 (1.23–3.20)
0.66 0.01
0.84 (0.61–1.15) 1.43 (0.85–2.42)
0.28 0.18
UO o0.1 ml/kg/h 3–6 h 6–12 h 12–24 h
3.05 (1.96–4.75) 5.79 (2.99–11.20) 9.44 (2.94–30.26)
o0.001 o0.001 o0.001
2.17 (1.35–3.48) 2.95 (1.20–7.25)d 5.74 (1.67–19.73)
o0.001 0.02 0.01
38/85 (44.7%) 23/38 (60.5%) 10/14 (71.4%)
Abbreviations: AKI, acute kidney injury; RRT, renal replacement therapy; UO, urine output. a Adjusted for: age, sex, Acute Physiology And Chronic Health Evaluation II diagnosis group, Simplified Acute Physiology Score II score without age and renal components, use of vasoactives, diuretics, and the cumulative fluid balance (% of baseline weight) on the day of oliguric episode. b Results are relative to those without oliguria or with oliguria lasting o3 h. c Includes interaction term between oliguria and use of vasoactives. d Includes interaction term of presence of oliguria and cumulative fluid balance.
OR of 2.36 (95% confidence interval 1.19–4.66), P = 0.01. The median (interquartile range) duration of UO o0.3 ml/kg/h among these patients was 7.1 (6.3–8.1) h and fluid accumulation on day 1 was 1.7 (0.2–4.0)% of baseline weight.
DISCUSSION
In this large prospective multicenter study among critically ill patients we analyzed the confounder-adjusted associations of consecutive oliguria of different severity and duration with the development of AKI defined by Cr criterion and/or RRT, and secondarily, with 90-day mortality. We found that episodes of severe oliguria (UOo0.1 ml/kg/h) for more than 4
3 h were independently associated with the development of AKI by Cr/RRT. The association of severe oliguria with the development of Cr-AKI makes physiologically sense. Oliguria is defined as UO o400 ml/d (0.2–0.3 ml/kg/h), which corresponds the maximally concentrated urine volume needed to excrete the solute load normally ingested.17 Critically ill patients receive large amounts of fluid and electrolytes during the resuscitation phase, which typically leads to fluid accumulation if kidneys are unable to respond. Most patients with AKI in the current study had fluid accumulation already on day 1. Co-existing fluid retention might obscure the detection of changes in serum Cr among oliguric patients.18 Moreover, Kidney International
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
90-Day mortality
Cr-AKI/RRT
100 90 80 70 60 50 40 30 20 10 0 None
170
0.5–3 h
3–6 h
<0.5 ml/ kg/h
0.3–0.5 ml/kg/h
731
411
Over 6 h
121
3–6 h
Over 6 h
3–6 h
0.1–0.3 ml/kg/h 342
91
6–12 h
12–24 h Over 24 h
<0.1 ml/kg/h 110
60
58
66
Figure 3 | 90-day mortality and presence of acute kidney injury defined by creatinine criterion and/or use of renal replacement therapy until day 5 in the ICU (% with 95% confidence intervals) according to the severity and duration of the longest oliguric episode of each patient. Numbers below each bar indicate the total number of patients in that group. The UO data of patients with RRT were screened also during RRT. Cr-AKI, acute kidney injury based on the Kidney Disease: Improving Global Outcomes creatinine criterion; RRT, renal replacement therapy; UO, urine output.
Table 4 | Univariate and multivariablea regression models for 90-day mortality Crude 90-day mortalityb
Univariate odds ratio (95% CI)c
Univariate P-value
Multivariable odds ratio (95% CI)c
Multivariable P-value
UO from 0.3 to o0.5 ml/kg/h 3–6 h 69/411 (16.8%) Over 6 h 31/121 (25.6%)
1.06 (0.078–1.45) 1.81 (1.16–2.83)
0.71 0.01
0.96 (0.68–1.37) 1.65 (1.00–2.72)
0.83 0.05
UO from 0.1 to o0.3 ml/kg/h 3–6 h 78/342 (22.8%) Over 6 h 26/91 (28.6%)
1.44 (1.08–1.92) 1.95 (1.21–3.14)
0.01 0.01
1.20 (0.87–1.65) 1.96 (1.13–3.38)
0.28 0.02
UO o0.1 ml/kg/h 3–6 h 34/110 (30.9%) 6–12 h 24/60 (40.0%) 12–24 h 31/58 (53.4%) Over 24 h 39/66 (59.1%)
1.95 2.91 5.01 6.30
(1.28–2.97) (1.71–4.94) (2.95–8.50) (3.81–10.43)
0.01 o0.001 o0.001 o0.001
2.08 3.04 6.78 9.76
(1.27–3.42) (1.50–6.15) (2.79–16.44) (4.07–23.43)
0.01 0.01 o0.001 o0.001
Abbreviations: RRT, renal replacement therapy; UO, urine output. a Adjusted for: age, sex, Acute Physiology And Chronic Health Evaluation II diagnosis group, Simplified Acute Physiology Score II score without age and renal components, use of vasoactives, diuretics, and RRT, and the cumulative fluid balance (% of baseline weight) on the day of oliguric episode. b Expressed as number/total number of patients with episode (%). c Results are relative to those without oliguria or with oliguria lasting o3 h.
because ICU patients generally do not receive full nutrition during their first days in the ICU, the amount of Cr to be excreted may be lower. These factors may explain why only severe oliguria as a sign of very low GFR was associated with AKI by Cr/RRT. We found the results regarding isolated oliguria interesting. First, patients with oliguric AKI but no Cr-AKI had Kidney International
higher crude mortality compared with those without any AKI. These results corroborate those of a database analysis reporting isolated stage 2–3 UO AKI to be associated with decreased age-adjusted one-year survival.11 Second, moderate isolated oliguria for over 6 h was independently associated with 90-day mortality. Notably, the median duration of oliguria among these patients was about 7 h. Patients 5
clinical investigation
ST Vaara et al.: Oliguria in the critically ill
With Cr-AKI/RRT
No Cr-AKI/RRT
80 70 60 50 40 30 20 10 0 None
0.5–3 h <0.5 ml/kg/ h
30 140
147 584
3–6 h
Over 6 h
0.3–0.5 ml/kg/h 82 329
29 92
3–6 h
Over 6 h
3–6 h
0.1–0.3 ml/kg/h 90 252
37 54
6–12 h
12–24 h
Over 24 h
<0.1 ml/kg/h 61 49
45 15
54 4
66 0
Figure 4 | Comparison of 90-day mortality (% with 95% confidence intervals) of patients who developed acute kidney injury diagnosed by creatinine-criterion and/or use of renal replacement therapy with those who did not. Data are presented according to the severity and duration of the longest oliguric episode of each patient. Numbers below each bar indicate the total number of patients in that group. Cr-AKI, acute kidney injury based on the Kidney Disease: Improving Global Outcomes creatinine criterion; RRT, renal replacement therapy.
presenting with severe isolated oliguria were few and the sample size may have been too small to show a significant association among them. Again, fluid accumulation may explain why these patients did not develop Cr-AKI. However, moderate isolated oliguria might be a sign of primarily non-renal derangements (e.g. hemorrhagic shock) that if resolved, may not ever lead to a rise in Cr or prolonged oliguria. The primary cause of these derangements may still expose the patients to adverse outcome. Thus, oliguria is an important warning sign regardless of later development of Cr-AKI. In our analysis, very few patients with severe oliguria did not develop Cr-AKI, but only a third of patients with mild or moderate oliguria developed Cr-AKI or received RRT. Notably, patients were diagnosed with AKI by Cr/RRT very early, after a median of 16 h in the ICU. Previously almost 50% of episodes of UO o0.5 ml/kg/h for over 4 h have been reported not to be followed by Cr-AKI.14 Although the incidence of Cr-AKI/RRT rose with increasing severity of oliguria, over a third of patients without oliguria developed Cr-AKI. Potentially, the etiology and clinical course for AKI may be different in Cr-AKI only patients compared with those presenting with oliguria also. Surprisingly, patients with Cr-AKI only did not have significantly higher crude 90-day mortality compared with patients without AKI of any type. This is in contrast with the findings by Kellum et al.11 However, the 90-day mortality of non-AKI patients in our analysis was higher (16.6%) compared with 7.3% reported previously,11 which is plausibly due to exclusion of patients 6
with scheduled post-surgery admission o24 h from the FINNAKI study. We defined the duration of oliguria as consecutive hours below the studied threshold according to the strictest interpretation of the UO criterion.4,5 The distinct ways of defining oliguria may explain the controversy of our results with a previous a study that did not find UO o0.5 ml/kg/h for 6 h to be associated with mortality after adjustments.16 In their single-center study comprising 725 patients they calculated UO in moving blocks during the first 24 h in the ICU and used the 6-hour period with lowest UO to define and stage AKI.16 Averaging the UO over a 6-hour collection period has been shown to be more sensitive than entailing consecutive oliguric hours but reduce the specificity of UO in predicting the development of AKI based on the AKI Network (AKIN) Cr criterion.8 To the best of our knowledge, this is the first prospective multicenter study to evaluate the association of hourly measured UO with a fixed mortality end-point. We were able to separate three different degrees of consecutive oliguria and study their independent association, first with development of AKI by Cr/, and second with 90-day mortality. Even mild oliguria for 6–12 h was independently associated with an increased risk for mortality, thus providing support for the adequacy of the current definition of oliguria. Moreover, our results corroborate those reported by Kellum et al.11 regarding that patients with both Cr-AKI and UO-AKI had the highest rates of RRT and 90-day mortality and that also isolated oliguria was associated with an adverse outcome. These Kidney International
ST Vaara et al.: Oliguria in the critically ill
findings underlie the importance of hourly UO measurements and necessity of using UO criterion along with Cr criterion to diagnose AKI. The strengths of our study include the prospective multicenter design, multiple data on confounding factors, and a fixed-time mortality. Several limitations, however, should be addressed. First, we did not record the volume status of oliguric patients at the start of the oliguric episode, and possibly, some of the patients may have been dehydrated. However, we adjusted for both daily fluid balance and use of diuretics in the multivariable models. Second, unmeasured confounders cannot be excluded despite that we adjusted for a number of important confounders. Third, we could not discern the potential effects of initiation of RRT on the UO but handled the data of RRT and non-RRT patients similarly regarding the mortality outcome. Fourth, as in previous studies on this topic,8,14,16 we excluded patients with an ICU length of stay shorter than 24 h to enable studying also longer periods of oliguria. Therefore, the overall mortality rate in this analysis was lower than in the main FINNAKI study.19 Fifth, among 80% of the patients, the baseline weight was reported by the patient or his/hers proxy or estimated by experienced personnel. We used the weight recorded at ICU admission as the baseline weight. No data which weight recording (baseline, actual, or ideal) would be the most accurate in assessing urine flow rate exist.6 Because most of the patients presented with their longest episode of oliguria on day 1 when the fluid accumulation was moderate in all patient groups (Table 1), we believe that using baseline weight instead of actual daily measured weight has not caused significant bias. Had we used the actual weight, the incidence of oliguria could be slightly greater. Baseline weight along with partly assessed patient weights were used also previously.16 Finally, the sample size of this study did not allow repeating the analyses in the subcohort of patients admitted with severe sepsis or in patients without diuretics. Despite these limitations, we consider the external validity of this study to be good among critically ill patients of Caucasian origin admitted to general ICUs. In conclusion, we demonstrated that episodes of severe oliguria (UO o0.1 ml/kg/h) for more than 3 h were independently associated with the development of Cr-AKI and/or administration of RRT. Patients with isolated oliguric AKI had higher crude mortality compared with patients without AKI. Our findings underlie the importance of hourly measurement of diuresis. MATERIALS AND METHODS This study was a secondary analysis using the prospectively collected data from the FINNAKI Study conducted in 17 Finnish ICUs between 1 September 2011 and 1 February 2012.19 The Ethics committee of Department of Surgery, Helsinki and Uusimaa Hospital District approved the study protocol and the use of deferred consent with written informed consent obtained from the patient or proxy as soon as possible. The Finnish National Kidney International
clinical investigation
Institute of Health approved collection of data from medical records of deceased patients if informed consent could not be obtained. The study was conducted according to the declaration of Helsinki. Patients All patients with either an emergency admission to ICU of any duration or post-surgical scheduled admission expected to exceed 24 h were eligible in the FINNAKI study if none of the following exclusion criteria were met: (1) pre-existing end-stage renal disease and need for maintenance dialysis, (2) previous receipt of RRT while enrolled in the study, (3) organ donor, (4) intermediate care patient, (5) patient not permanently living in Finland, (6) patients transferred from another participating ICU after full study data collection period, and (7) absence of consent. From the current analysis, we further excluded several patient groups presented in the study flow chart (Figure 1). Data collection We recorded data from patients’ pre-existing conditions, medications, and ICU treatment with case report forms and obtained physiological data including UO, laboratory values, diagnoses, and severity of illness scores from the ICU data management system via the Finnish Intensive Care Consortium platform. The data collection period was until day 5 in the ICU. All patients had urine catheters and UO was registered hourly in the electronic patient charts. Patient weight recording at ICU admission in all participating centers was based either on weighting (in the ICU or preoperatively in 426 patients (19.7%)) or assessed (according to the reported weight by the patient or his/her proxy or estimated bedside by personnel) in 1734 (80.3%). The baseline weight determined at ICU admission was used throughout the study to calculate urinary flow rates. The intensivists screened patients daily for the presence of severe sepsis according to the ACCP/SCCM definition.20 We used the KDIGO criteria to screen for and stage AKI.6 Plasma Cr measurements were performed in accredited hospital laboratories using enzymatic method with an interassay coefficient of variation o3–4%. As baseline Cr, we used the latest value obtained within 1 year but at least a week before study inclusion. If baseline Cr value could not be obtained, we estimated it using the Modification of Diet in Renal Disease equation assuming a GFR of 75 ml/min/1.73 m2 as recommended.4 Oliguria An electronic calculator was constructed to calculate the UO ml/kg/h between two UO recordings. We screened these data for the presence of the following degrees of severity oliguria: (1) o0.5 ml/kg/h, (2) o0.3 ml/kg/h, and (3) o0.1 ml/kg/h. Regarding the development of AKI by Cr/RRT, we screened the data before fulfillment of this endpoint. Regarding 90-day mortality, we screened data until ICU discharge or day 5 in the ICU, and included UO data obtained during RRT among those with RRT. We recorded the duration of each episode of consecutive values below the threshold. We formed exclusive categories of the severity of oliguria (mild oliguria from 0.3 to o0.5 ml/kg/h, moderate oliguria from 0.1 to o0.3 ml/kg/h) by classifying the patients according to their longest oliguric episode without periods of deeper oliguria more than 3 h. Definition of severe oliguria was UO o0.1 ml/kg/h. The duration of episodes were categorized into o3 h, 3–6 h, 6–12 h, 12–24 h, and longer than 24 h. 7
clinical investigation
Association of oliguria with the endpoints We used logistic regression to study the association between the different types of oliguric episodes with the development of AKI by Cr/RRT, and secondarily with 90-day mortality. We adjusted for the following a priori selected confounders: (1) fluid accumulation (%) of baseline weight, (2) use of diuretics, (3) use of vasoactive medication, (4) SAPS II score (day 1) without points given for age and renal components, (5) age, (6) sex, and (7) APACHE II admission diagnosis group. For confounders 1–3, we used values obtained on the day of the beginning of the oliguric episode. Regarding 90-day mortality, we also adjusted for the use of RRT. We calculated the median day number of occurrence of each type of oliguric episode, and used data on that day for adjustments among patients without the tested type of oliguric episode. We used multiplicative terms to test for presence of interactions between confounders 1 and 3 and use of RRT with oliguria. We assessed the model fit with the Hosmer–Lemeshow test.
ST Vaara et al.: Oliguria in the critically ill
6.
7. 8. 9.
10.
11. 12.
13.
14.
Statistical analysis We report categorical variables with count and percentage and continuous variables with median and interquartile range. We used Fischer’s exact test to compare categorical variables and Mann–Whitney U-test to compare continuous variables, or in case of comparing groups of three or more χ2 test or Kruskal–Wallis test. We used logistic regression with enter method to adjust for confounders. We set the significance level at 0.05 and report two-tailed P-values. We conducted the analysis with SPSS statistics 20.0. (IBM, Armonk, NY, USA). DISCLOSURE
All the authors declared no competing interests. ACKNOWLEDGMENTS
We are grateful to professor Seppo Sarna for valuable statistical advice. We thank Tieto Healthcare & Welfare for database management. The Academy of Finland, the Sigrid Juselius Foundation, Päivikki and Sakari Sohlberg Foundation, and Helsinki University Central Hospital EVO-grants have supported this study. SUPPLEMENTARY MATERIAL Figure S1. The 90-day mortality (%, 95% confidence interval) according to the longest oliguric episode of patients who did not require RRT. Supplementary material is linked to the online version of the paper at http://www.nature.com/ki. REFERENCES 1.
2.
3.
4.
5.
8
Uchino S, Kellum JA, Bellomo R et al. Acute renal failure in critically ill patients: a multinational, multicenter study. JAMA 2005; 294: 813–818. Chertow GM, Burdick E, Honour M et al. Acute kidney injury, mortality, length of stay, and costs in hospitalized patients. J Am Soc Nephrol 2005; 16: 3365–3370. Joannidis M, Metnitz B, Bauer P et al. Acute kidney injury in critically ill patients classified by AKIN versus RIFLE using the SAPS 3 database. Intensive Care Med 2009; 35: 1692–1702. Bellomo R, Ronco C, Kellum JA et al. Acute renal failure - definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004; 8: R204–R212. Mehta RL, Kellum JA, Shah SV et al. Acute Kidney Injury Network: report of an initiative to improve outcomes in acute kidney injury. Crit Care 2007; 11: R31.
15.
16. 17. 18.
19.
20.
Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group: KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl 2012; 2: 1–138. Ricci Z, Cruz D, Ronco C. The RIFLE criteria and mortality in acute kidney injury: a systematic review. Kidney Int 2008; 73: 538–546. Macedo E, Malhotra R, Bouchard J et al. Oliguria is an early predictor of higher mortality in critically ill patients. Kidney Int 2011; 80: 760–767. McIlroy DR, Argenziano M, Farkas D et al. Incorporating oliguria into the diagnostic criteria for acute kidney injury after on-pump cardiac surgery: impact on incidence and outcomes. J Cardiothorac Vasc Anesth 2013; 27: 1145–1152. Wlodzimirow KA, Abu-Hanna A, Slabbekoorn M et al. A comparison of RIFLE with and without urine output criteria for acute kidney injury in critically ill patients. Crit Care 2012; 16: R200. Kellum JA, Sileanu FE, Murugan R et al. Classifying AKI by Urine Output versus SerumCreatinine Level. J Am Soc Nephrol 2015. Mandelbaum T, Lee J, Scott DJ et al. Empirical relationships among oliguria, creatinine, mortality, and renal replacement therapy in the critically ill. Intensive Care Med 2013; 39: 414–419. Mandelbaum T, Scott DJ, Lee J et al. Outcome of critically ill patients with acute kidney injury using the Acute Kidney Injury Network criteria. Crit Care Med 2011; 39: 2659–2664. Prowle JR, Liu YL, Licari E et al. Oliguria as predictive biomarker of acute kidney injury in critically ill patients. Crit Care 2011; 15: R172. Leedahl DD, Frazee EN, Schramm GE et al. Derivation of urine output thresholds that identify a very high risk of AKI in patients with septic shock. Clin J Am Soc Nephrol 2014; 9: 1168–1174. Md Ralib A, Pickering JW, Shaw GM et al. The urine output definition of acute kidney injury is too liberal. Crit Care 2013; 17: R112. Klahr S, Miller SB. Acute oliguria. N Engl J Med 1998; 338: 671–675. Liu KD, Thompson BT, Ancukiewicz M et al. Acute kidney injury in patients with acute lung injury: impact of fluid accumulation on classification of acute kidney injury and associated outcomes. Crit Care Med 2011; 39: 2665–2671. Nisula S, Kaukonen KM, Vaara ST et al. Incidence, risk factors and 90-day mortality of patients with acute kidney injury in Finnish intensive care units: the FINNAKI study. Intensive Care Med 2013; 39: 420–428. Bone RC, Balk RA, Cerra FB et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/ SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992; 101: 1644–1655.
FINNAKI STUDY GROUP
Central Finland Central Hospital: Raili Laru-Sompa, Anni Pulkkinen, Minna Saarelainen, Mikko Reilama, Sinikka Tolmunen, Ulla Rantalainen, Marja Miettinen East Savo Central Hospital: Markku Suvela, Katrine Pesola, Pekka Saastamoinen, Sirpa Kauppinen Helsinki University Central Hospital: Ville Pettilä, Kirsi-Maija Kaukonen, Anna-Maija Korhonen, Sara Nisula, Suvi Vaara, Raili Suojaranta-Ylinen, Leena Mildh, Mikko Haapio, Laura Nurminen, Sari Sutinen, Leena Pettilä, Helinä Laitinen, Heidi Syrjä, Kirsi Henttonen, Elina Lappi, Hillevi Boman Jorvi Central Hospital: Tero Varpula, Päivi Porkka, Mirka Sivula, Mira Rahkonen, Anne Tsurkka, Taina Nieminen, Niina Pirttinen Kanta-Häme Central hospital: Ari Alaspää, Ville Salanto, Hanna Juntunen, Teija Sanisalo Kuopio University Hospital: Ilkka Parviainen, Ari Uusaro, Esko Ruokonen, Stepani Bendel, Niina Rissanen, Maarit Lång, Sari Rahikainen, Saija Rissanen, Merja Ahonen, Elina Halonen, Eija Vaskelainen Lapland Central Hospital: Meri Poukkanen, Esa Lintula, Sirpa Suominen Länsi-Pohja Central Hospital: Jorma Heikkinen, Timo Lavander, Kirsi Heinonen, Anne-Mari Juopperi Middle Ostrobothnia Central Hospital: Tadeusz Kaminski, Fiia Gäddnäs, Tuija Kuusela, Jane Roiko North Karelia Central Hospital: Sari Karlsson, Matti Reinikainen, Tero Surakka, Helena Jyrkönen, Tanja Kidney International
ST Vaara et al.: Oliguria in the critically ill
Eiserbeck, Jaana Kallinen Oulu University Hospital: Tero AlaKokko, Jouko Laurila, Sinikka Sälkiö Satakunta Hospital District: Vesa Lund, Päivi Tuominen, Pauliina Perkola, Riikka Tuominen, Marika Hietaranta, Satu Johansson South Karelia Central Hospital: Seppo Hovilehto, Anne Kirsi, Pekka Tiainen, Tuija Myllärinen, Pirjo Leino, Anne Toropainen
Kidney International
clinical investigation
Tampere University Hospital: Anne Kuitunen, Jyrki Tenhunen, Ilona Leppänen, Markus Levoranta, Sanna Hoppu, Jukka Sauranen, Atte Kukkurainen, Samuli Kortelainen, Simo Varila Turku University Hospital: Outi Inkinen, Niina Koivuviita, Jutta Kotamäki, Anu Laine Vaasa Central Hospital: SimoPekka Koivisto, Raku Hautamäki, Maria Skinnar.
9