Virus infection can lead to a wide range of kidney diseases through the heterogeneous mechanisms, and virus-associated AKI is well-known phenomenon [10]. Major epidemics such as Severe acute respiratory syndrome (SARS) and Middle East Respiratory Syndrome coronavirus (MERS-COV) mainly affect the respiratory system. Interestingly, the aforementioned infections commonly afflict the kidney and have been associated with long-term adverse renal outcomes. A retrospective study by Kwok et al. demonstrated that in 2003, the incidence rate of AKI was 6.7% in SARS infection and was associated with a higher mortality rate of 91.7% [11]. In addition, during the outbreak of Middle East Respiratory Syndrome coronavirus (MERS-COV), up to 58% of the patients admitted to the intensive care unit needed renal replacement therapy for AKI in Saudi Arabia [12].

The incidence of AKI due to COVID-19 has been reported to vary worldwide, ranging from 5%- 56% [13, 14]. Meta-analysis revealed that the observed difference might be due to geographical variation and the proportion of severely ill patients reported in each study. Western countries reported a higher incidence of AKI in COVID-19 infection than Asian countries. An observation study in New York reported 46% of AKI among the hospitalized COVID-19 infection, with the proportion of white, black, and Asian being 24%, 29%, and 4%, respectively [15]. Jewell et al. revealed that 39% of hospitalized COVID-19 developed AKI, with blacks having the highest proportion rate (46%) followed by whites (36%) and Asians (32%) in London [16]. In contrast, another retrospective study in the United Kingdom reported a lower rate of AKI (20.3%) among hospitalized COVID-19 patients, but the majority of the admissions were of white ethnicity, with 70.3% white, 9.1% black and 11% Asian, that’s markedly different from the Jewell et al. cohort [17]. Thus, the incidence rate of AKI in Western countries seemed to depend on the study's proportion of black ethnicity. Of note, Asian ethnicity tends to have a lower incidence rate of AKI in Western countries. This is further supported by a subgroup analysis of a systemic review that reported a two-fold rise in the incidence of AKI in non-Asian populations, especially those of African ancestry, significantly more than the Asian population [18]. Our Asian cohort mainly consists of Malay, Chinese, and Indian ethnicities without black or African ancestry and hence explaining the lower incidence of AKI. In our study, we noted 323 out of 1529 patients (21.1%) developed AKI, which is lower than the reported incidence rate from Western countries.

In our study, the classification of AKI severity is based on KDIGO criteria 2012. We noted that most hospitalized patients who developed AKI predominantly fell into stage 1 (77%), which is in keeping with findings from most other studies indicating stage 1 AKI was the commonest occurrence [14, 19,20,21]. Severe AKI (stage 2 or 3) was more prevailing in critically ill patients who needed intensive care [22, 23] and it was associated with higher mortality. Indeed, the risk of mortality increases with the severity of AKI in COVID-19 patients, as demonstrated by ISARIC WHO CCP-UK Study which showed that AKI stage 1 has an odd ratio (OR) of 1.58 (95% CI 1.49- 1.67) on 28-day mortality, and the OR rose further to 3.50 (95% CI 3.14–3.91) for AKI stage 3 [20]. In our study, a 34.7% mortality rate was observed in those who had AKI across all spectrums of stages. Our findings are consistent with the reported mortality rate in the study by Morieri M et al., which demonstrated a mortality rate of 39.3% in AKI patients, confirming that AKI was an independent predictor for mortality after adjusting for cofounders [24]. The noted poor outcome in our analysis may be due to the difficulty in caring for an overwhelming number of COVID-19 patients in a single center with limited resources such as beds, staffing, and equipment for renal replacement therapy during the peak of the COVID-19 pandemic.

Urinary abnormalities, including proteinuria, hematuria, and leukocyturia, are seen in hospitalized COVID-19 patients. In our study, not all the patients had urinalysis; however, among those who had urine examination, approximately 56% of the non-AKI patients had either proteinuria or microscopic hematuria, leukocyturia, whereas a higher (92%) of AKI COVID-19 patients had mentioned urinary disorders, thus concurring with Chan et al. who reported a greater proportion of urinalysis abnormalities in patients with AKI, as the author reported 80% had either proteinuria or hematuria, and 60% had leukocyturia [15]. Thus, this highlights the importance of urinalysis for hospitalized COVID-19 patients, as the presence of proteinuria/microscopic hematuria has been reported to be an independent risk factor for AKI, morbidity, and mortality [13].

In light of AKI and urinary abnormalities on adverse patients’ outcome, the importance of understanding the mechanism of AKI in COVID-19 has been prioritized and intensively studied, with the intention of devising a proper preventive measure and treatment strategies. Kidney disease among COVID-19 patients may manifest with different clinical phenotypes. The consensus report of the 25th Acute Disease Quality Initiative (ADQI) work group postulated different mechanisms through which COVID-19 contributes to AKI [25]. The direct effect of viruses resulting in collapsing glomerulopathy in renal has been well-reported in case studies [26,27,28,29]. Multisystem inflammatory syndrome (MIS-A) related to COVID-19 infection has been described [30], postulating hyperinflammation, uncontrolled complement activation, and endothelial dysfunction having direct insult to the kidney structures. A review article by Sharma P et al. delineates the pathological changes of COVID-AKI due to various mechanisms with acute tubular injury, collapsing glomerulopathy, and thrombotic microangiopathy frequently seen in both living and autopsied renal tissues. Other less common findings noted were podocytopathies, lupus nephritis, anti-glomerular basement membrane disease and anti-neutrophil cytoplasmic antibody vasculitis [31]. Most importantly, indirect injury to the renal system, e.g., hypovolemia, sepsis-associated AKI, acute tubular necrosis, acute interstitial nephritis due to nephrotoxic medications, and increased positive end-expiratory pressure ventilator induce renal angiotensin activation aldosterone system, is thought to be the major culprits.

Our cohort revealed that severe AKI requiring dialysis during hospitalization was low (4.6%) among the AKI patients (15 out of 323 patients); however, more than one-third of the patients had no renal recovery during the follow-up review. A study in New York City showed that, overall, 19% of COVID-19 patients needed RRT, and a third of patients had no renal recovery from COVID-19 [15]. In contrast, a prospective observational study of 4613 patients with COVID-19 AKI in India reported a high RRT rate of 40.5% with a high non-recovery rate of 72% upon discharge till three months follow-up, and 49% of the patients eventually progressed to chronic kidney disease [32]. The disparity in dialysis requirement and the rate of renal recovery in various centers could be attributed to several factors. The most likely cause is the different thresholds of initiating RRT in different centers. Centers with lower thresholds for dialysis may potentially have worse renal outcomes. This is supported by a retrospective study illustrating the rate of renal recovery in severe AKI needing dialysis had a renal recovery rate of 61% on days 1–4 and 8% in 31–90 days [33]; Other factors may be related to the variation of the duration of the post-hospitalization follow-up and definition of renal recovery thus leading to the disparity in the reported results. This emphasizes the importance of understanding the pathophysiology and risk factors of COVID-related AKI and the clinical progress post-COVID-19 recovery. The association of non-recovery AKI and evolution to chronic kidney disease is well-known in other diseases [34, 35]; however, whether a similar theory applies to COVID-19 AKI is uncertain as of now. In the event of a high rate (≈30%) of non-renal recovery COVID-19 AKI, we foresee the clinical challenges of dealing with an increasing chronic kidney disease population. Thus, serial monitoring of the renal function in hospitalized COVID-19 patients is crucial, and those with AKI will require a longer period of review to determine renal recovery. This will allow the implementation of preemptive measures involving the prevention and retardation of chronic kidney disease.

With the aforementioned high proportion of non-recovery AKI in COVID-19 underscores the importance of early prediction of the clinical phenotype by analyzing factors associated with AKI hospitalized COVID-19. In this study, we dissected the risk factors based on demographic and clinical parameters and found that age and pre-existing CKD were statistically significant in predictors of AKI in hospitalized COVID. These findings concur with other reports indicating age as an independent predictor for AKI [14, 36, 37]. Furthermore, age has also been reported as a risk factor for death in COVID-19 patients [19, 38], suggesting that elderly population is at a higher risk of morbidity and mortality, warranting special attention upon admission. In addition, similar to our study cohort, a retrospective study in London demonstrated a three-fold risk of AKI for patients who developed COVID-19 AKI with pre-existing CKD [16]. Diabetes and hypertension is highly prevalent worldwide [39, 40]. Patients who contracted COVID-19 with underlying comorbidities were associated with poorer outcomes [41, 42]. In our cohort, patients with either diabetes or hypertension were independent predictors for AKI, and our findings were supported by other studies [43, 44]. In addition, ferritin, an acute phase reactant, tends to be elevated in inflammatory conditions, including infections. Reports have shown that a raised ferritin level can predict AKI among COVID-19 patients, and this is consistent with our cohort [45, 46]. The ROCs of serum ferritin in predicting AKI are shown in Fig. 3 with an AUC of 0.711, representing a moderate ability to discriminate prediction for the development of AKI. An optimal cutoff point of 500 ug/l in our analysis allows a performance characteristic of a positive predictive value [PPV] of 34% and a negative predictive value [NPV] of 87%. Our study did not track the ferritin trend during hospitalization, unlike Jonathan Feld et al., who demonstrated ferritin levels at different time frames may be useful as a predictor for the need for renal replacement therapy (RRT) among hospitalized COVID-19 patients. His study reported the presentation of ferritin with a threshold above 569 ng/ml (sensitivity of 0.84 and specificity of 0.46) and maximum ferritin of more than 2365 ng/ml (sensitivity of 0.62 and specificity of 0.74) are useful as a predictors for severe AKI requiring RRT [47]. Thrombocytopenia is a well-known clinical sequela in viral infection [48]. However, the reported incidence rates were less in COVID-19 patients ranged from 5–41.7% [49, 50]. The seemingly lower number may be due to the distinct clinical phenotype of COVID-19 to other types of viral infection as it is associated with hypercoagulable state. There are various possible mechanisms of COVID-19-associated thrombocytopenia, including platelet activation, platelet clearance, platelet autoantibody formation, splenic sequestration, and marrow suppression [51]. Our study showed that the incidence rate of thrombocytopenia was two times higher in the AKI group (n = 73; 22.6%) than AKI-naïve group (n = 126; 10.4%); p < 0.001 even though the median platelet counts were within the normal range for both groups. The apparent normal platelet count may be explained by a corresponding increase in platelet production despite an increase platelet consumption and hyperreactivity. This finding concurs with Taha et al., that reported increased platelet activation in COVID-19 patients as measured by a higher mean platelet volume despite a normal platelet count. In his study, there was a high incidence of AKI in critically ill patients with unique platelet features, that including larger, more granular platelets and a higher mean platelet volume [52]. We found that platelet counts were an independent predictor for AKI in hospitalized COVID-19 patients.

Most experts think that vaccine is the most effective way of curbing the global spread of viral disease that including Covid-19 infection. Intensive research was undertaken in the early phase of the pandemic, and various types of COVID-19 vaccines were being developed and listed for emergency use by WHO, which comprised mRNA vaccine (Pfizer BioNTect), whole virus vaccine (Sinovac) and non-replicating viral vector (Oxford-AstraZeneca). Our study revealed that Pfizer vaccination had a significantly positive renal outcome, with the majority of the vaccinated patient (90.6%) not developing AKI. Patients who received either AstraZeneca (18.5%) or Sinovac (21.4%) vaccines also had a lower risk of developing in hospitalized COVID-AKI, albeit it was not statistically significant in our study. We believe that the immunogenicity of various types of COVID vaccines may confer a different protective effect from acute kidney injury; however, this will require further analysis.

Despite the perceived positive effect of the COVID-19 vaccine on renal outcome in our cohort, there is a potential pitfall with the review article by Yebei Li which highlighted 52 cases of Acute Kidney Disease (AKD) that encompassing podocytopathy, IgA nephropathy, vasculitis, tubular interstitial nephritis and thrombotic microangiopathy following the Covid vaccination [53] The causal relationship between COVID-19 vaccination and AKI is uncertain for now, however considering the low rate of vaccine-induce AKI, coupled with the significant risk reduction of adverse patient outcomes related to COVID-19 infection, it would be sensible to implement vaccination in a population but with the understanding of the potential side effect of the vaccines.

There are a few limitations in our study. Firstly, it is a retrospective study, and not all the patients have the baseline serum creatinine. Hence, we used the on-arrival serum creatinine at the emergency department for those who lacking baseline serum creatinine. In our study, the timing of post-discharge serum creatinine varied substantially, ranging from two weeks up to three months after being discharged from the ward. Therefore, we may prematurely label some patients as having no renal recovery if the latest post-discharge serum creatinine was obtained before three months after the AKI episode. Furthermore, the timing of AKI development during hospitalization was not consistently recorded, even though the author believe it is important to look for the temporal relationship between the onset of AKI and the severity of COVID-19 infection. The diagnoses of pneumonia and acute respiratory distress syndrome (ARDS) were based on the treating physician or anesthetist assessment rather than interpretation by an expert such as a radiologist or pulmonologist. Despite these limitations, a large sample size of more than a thousand patients in this study provides robust data reflecting the Malaysian population.

In conclusion, the incidence of AKI is not rare among hospitalized COVID-19 patients and is associated with prolonged hospitalization, poorer renal outcomes, and higher mortality. Our study suggests that COVID-19 vaccination reduces the incidence rate of AKI, but future studies are warranted to look for the long-term AKI outcome in the COVID-19 vaccinated patients. Given the adverse impact of AKI on COVID-19 patients, earlier detection of predictors, including the demographic factors and clinical parameters, is essential to improve the patient renal and survival outcomes.