This study is, to our knowledge, the first to characterize and analyze the prognostic impact of AKI according to its duration in severe COVID-19 patients. In this observational study, we illustrated that 45% (48/107) of severe patients with COVID-19 developed AKI, including 22.9% (11/48) patients with transient AKI and 77.1% (37/48) patients with persistent AKI.
AKI is common among severe patients with COVID-19, affecting approximately 26–76% of patients admitted to the ICU[6–8]. In-hospital mortality rates of AKI range from 50% overall to 42% in the ICU.[7] AKI could be distinguishable in terms of renal recovery from AKI or the duration of AKI.[25] Recent findings have led to the hypothesis that transient AKI and persistent AKI in critically ill patients might share similar pathophysiological mechanisms and that the duration of AKI might reflect its severity rather than its mechanism.[26] In the present study, patients with persistent AKI had significantly higher death rates than those transient AKI. Although exact pathophysiology is not clearly elucidated in, experimental models have shown that even with apparent renal function recovery from AKI, histologic and physiologic changes may persist after AKI. Recently, the 2017 Acute Disease Quality Initiative (ADQI) 16 workgroup defined persistent AKI as the continuance of AKI diagnosed by the Scr level or urine output criteria for > 48 hours after AKI onset. So early recognition of transient AKI secondary to severe COVID-19 and use of supportive and therapeutic measures to avoid further kidney damage are crucial to reduce morbidity and mortality. Several studies have examined the prognostic impact of AKI duration using different diagnostic criteria as follows: transient AKI (≤ 7 days) and persistent AKI (> 7 days) using the KDIGO criteria;[27, 28] transient AKI (Scr ≤ 115 µmol/L at discharge) and persistent AKI (Scr > 115 µmol/L at discharge) using the KDIGO criteria;[29] short (≤ 2 days), medium (3–5 days), and long (≥ 6 days) using the KDIGO criteria;[15] transient AKI (≤ 3 days) and persistent AKI (> 3 days) using the AKIN or KDIGO criteria.[16, 17] The use of a universally recognized definition of AKI would improve our knowledge regarding practices for, research on, and public health issues related to AKI.
Previous studies on the prognostic impact of AKI duration yielded conflicting results. For example, Mizota and colleagues published a retrospective cohort study involving 258 AKI patients (median age 66 years) undergoing major abdominal surgery.[28] Using the KDIGO criteria, the postoperative AKI patients were divided into transient AKI and persistent AKI groups based on the time when the Scr level returned to the no-AKI range within 7 days after surgery. The authors reported that both transient AKI and persistent AKI were independently associated with 1-year mortality. Also, the mortality rate was higher in the persistent AKI than the transient AKI group. However, another retrospective study found no significant association between AKI duration and adverse outcomes.[17] The discrepancies between these studies might reflect differences in the duration of recovery from AKI according to the definition of AKI, transient AKI, or persistent AKI; the participants’ age; and the etiology of AKI. One of the limitations is that we not further classified persistent AKI into categories based on AKI duration.
In the present study, we found that patients with both transient AKI and persistent AKI had significantly higher death rates than those without AKI in elderly patients with severe COVID-19. The ADQI 16 workgroup defined transient AKI as Scr that returned to baseline within 48 h after AKI onset calling for at least two measurements over a 48-hour period and baseline Scr available prior to admission. However, this can be difficult in clinical practice with emerging infectious diseases. In addition, previous publications from developed countries report the proportion of patients with more than 2 Scr measurements during hospitalization ranged from 63.2–67.6%, which is much higher than the figure reported in China only 25–30%.[30–32] Therefore, some patients with transient AKI could be misclassified as not having AKI.[33] Indeed, improving survival only via recognized AKI is not enough. It is, for instance, absence of kidney care is also a negative factor affecting the outcomes. In 2009, the National Confidential Enquiry into Patient Outcome and Death reported up to 50% of patients who died from AKI had not received ‘good’ kidney care; in addition, for 20% of these patients, the cause was both predictable and preventable.[34] Thus, increasing the likelihood of a timely AKI diagnosis and identifying patients who need kidney care to preventing the progression of transient AKI to persistent AKI remains a challenge.
The classification of AKI by duration may discriminate between patients with transient AKI (pre-renal) or hemodynamic AKI that does not involve any true injury to the renal tubular cells and those with true intrinsic AKI (that is, structural kidney injury). Furthermore, the duration of AKI may be a surrogate of the renal recovery potential of the injured kidney or continued on-going insults. The parameters used for evaluating patients’ renal function were glomerular filtration rate (GFR) or Scr. The recent definition from the KDIGO guideline classifies AKI by increasing severity, from stage 1 to stage 3, based on the Scr increase and/or decrease in urine output, and more severe AKI stages are associated with adverse outcomes.[10, 11] However, we found no association between in-hospital mortality and AKI severity in this study. It has been demonstrated that in elderly patients, renal recovery may be prolonged and frequently incomplete. Additionally, the duration of AKI may likely denote the overall illness severity of the patient, as those who are more severely ill and have continued extrarenal organ dysfunction will take longer to recover. Therefore, the KDIGO diagnosis and staging of AKI with Scr may not be suitable for the elderly population with severe COVID-19. Herein, we demonstrated that the duration of AKI still provided prognostic information over the KDIGO stage alone and can provide additive risk information for mortality risks for elderly patients, especially when AKI duration is longer than 48 hours.
The study had the following limitations. (1) First, medical data of this study were collected from a single-center, and the number of patients included is limited; thus, the results may lack generalizability. (2) Second, because of the strain on medical resources, urine output data of most patients were missing and not been collected, which is one component of the AKI definition; therefore, the incidence of AKI may be underestimated. (3) Due to the limited number of patients included, AKI patients were not further divided into categories based on AKI duration, such as short duration: resolving AKI lasting 3–4 days; medium duration: resolving AKI lasting 5–7 days; and long duration: AKI lasting > 7 days. Thus, a larger AKI patient population is needed to confirm these results. (4) Finally, we lacked data after the patients were discharged.