1 Introduction

Sepsis is a common life-threatening clinical syndrome in the intensive care units, affecting 20–30 million people worldwide each year. Severe sepsis can directly lead to multi-organ dysfunction and is the leading cause of death in ICU patients. Acute kidney injury (AKI) is a common and serious complication of sepsis. The incidence rate of Sepsis associated acute kidney injury (SAKI) is up to 47.5% [1].

SAKI is an acute organ dysfunction syndrome and is a common complication in critically ill patients. There are approximately 6 million cases worldwide each year [2]. SAKI is associated with long hospital stays, high mortality rates and high medical costs in ICU patients. It has imposed a huge economic burden on society. The pathogenesis of SAKI includes hemodynamic disorder, endothelial cell damaged by cytokines and oxygen free radicals, abnormal coagulation and energy metabolism. However, the pathogenesis of SAKI has not been fully elucidated [3]. Treatment for SAKI includes early fluid resuscitation, antibiotics and renal replacement. The pathogenesis of SAKI is complexity and early treatment is important, which indicates that early diagnosis of SAKI by clinicians is essential. Up to now, doctors generally judge whether patients have AKI by indicators such as serum creatinine and urine volume. However, serum creatinine and urine volume have some limitations in the diagnosis of AKI—elevated serum creatinine concentration is an insensitive indicator of kidney injury and is limited by the lack of baseline values in many patients [4]. Oliguria may be nonspecific in SAKI. This meta-analysis aims to analyze the risk factors for the occurrence of SAKI, so as to help the early prediction of SAKI.

2 Methods

2.1 Inclusion Criteria

Patients that met the following criteria were included: A. Meeting the diagnostic criteria for sepsis or septic shock (sepsis 1.0, sepsis 2.0, sepsis 3.0). B. Meeting the diagnostic criteria for AKI (KDIGO, AKIN, RIFLE). C. Original data provides OR value and 95%CI or data can be converted to OR value and 95% CI.

2.2 Exclusion Criteria

A. Pre-existing renal insufficiency or other renal disease. B. AKI not due to sepsis (eg. renal transplantation, use of nephrotoxic drugs or contrast agents, urinary obstruction). C. No risk factors in the outcome index. D. NOS quality score < 6. E. Repeated and reviewed literature.

2.3 Data Extraction

Taking Pubmed as an example, the search strategy is: ((Sepsis [MeSH]) OR (sepsis[Title/Abstract]) OR (infective shock [Title/Abstract]) OR (bloodstream infection[Title/Abstract]) OR (septic shock [Title/Abstract]) OR (pyemia [Title/Abstract]) OR (pyohemia [Title/Abstract]) OR (severe sepsis [Title/Abstract])) AND ((acute kidney injury [MeSH]) OR (acute kidney injury [Title/Abstract]) OR (acute renal injury [Title/Abstract]) OR (acute kidney failure [Title/Abstract]) OR (acute renal failure[Title/Abstract]) OR (acute renal insufficiencies[Title/Abstract])) AND ((risk factor) OR (factor) OR (influence factor) OR (relevant factor) OR (hazard)).

Firstly, the titles and abstracts of all retrieved articles were read to exclude irrelevant articles. Secondly, Read the full text, articles that did not meet the criteria were excluded according to the inclusion and exclusion criteria (Fig. 1). Finally, data were extracted from the included studies, including author's name, country or region, year of publication, research design, sample size, data of SAKI patients and other information (Table 1). Extract study results including relevant risk factors, OR values, 95% CI and P values.

Fig. 1
figure 1

Flow diagram of search strategy and study selection

Full size image
Table 1 Information and quality evaluation of included literature
Full size table

2.4 Quality Analysis

All included studies were examined by the NOS quality assessment scale, which is available at http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. NOS score ≥ 6 was considered high-quality articles and were included in the study, and exclude the low-quality articles.

2.5 Statistical Analysis

After extracting data of the 25 included studies, the OR values and 95% CI of risk factors were imported into Stata SE 16.0 software. Heterogeneity test was performed on the study data first. If I2 ≥ 50%, inter-group data were considered to be heterogeneous, the combined amount was calculated using random effect model. If I2 < 50%, the fixed-effect model is adopted (Table 2). Pooled OR and 95%CI were finally obtained. P ≤ 0.05 were considered to be statistically significant. The relevant data of each risk factor obtained were summarized and drawn the forest map (Fig. 2).

Table 2 Heterogeneity and effects model
Full size table
Fig. 2
figure 2

Forest plot of meta-analysis

Full size image

3 Result

3.1 Literature Search Process and Results

By 2022.02 we retrieved 3,323 studies from Pubmed, Web of Science, Cochrane library, Wiley online library and CNKI databases. After reading the titles and abstracts, excluding irrelevant and duplicate studies, we included 51 studies. We finally included 25 studies in 8 countries by reading the full text.

3.2 Basic Information of the Included Literature and Quality Evaluation Results

There were 8 case–control studies, 12 retrospective cohort studies, 4 prospective cohort studies, and 1 RCT study among the 23 included papers. The quality of all included study articles was good according to the NOS quality evaluation scale.

3.3 Results of Meta-Analysis of Risk Factors in the Literature Related to SAKI

Meta-analysis of 13 related risk factors for SAKI (including general factors, comorbidities, and laboratory examination results) revealed statistically significant risk factors (P < 0.05) are as follow: diabetes, hypotension(MAP < 65 mmHg), coagulation dysfunction, liver disease, heart disease, high APACHEII score, lactic acidosis, oliguria, high serum creatinine (> 88 mol/l) (Table 3).

Table 3 Meta-analysis of risk factors of SAKI
Full size table

3.4 Sensitivity Analysis

Sensitivity analyses were performed by changing the analysis model and re-combining the statistics after excluding each study in turn, and none of the results changed from one side to the other, suggesting more stable results. Sensitivity analysis of the study results was performed by changing the analysis model and excluding literature with large sources of heterogeneity or weights to observe the stability of the study results.

3.5 Publish Bias Analysis

Funnel plots were drawn for risk factors (diabetes mellitus, lactic acidosis) with ≥ 7 studies, we found the funnel plots are asymmetrical. The result suggests that publication bias exists in the theses two risk factors (Fig. 3, 4).

Fig. 3
figure 3

Funnel plot of Diabetes

Full size image
Fig. 4
figure 4

Funnel plot of Lactic acidosis

Full size image

4 Discussion

SAKI, a kind of renal hypoperfusion and inflammatory factor response storm caused by sepsis or septic shock, further causes necrosis of renal tubular epithelial cells, resulting in acute renal function damage. The long hospital stay, high mortality rate and high medical expenses of this disease have imposed a huge economic burden on society and families, arousing extensive concern from medical researchers. The pathogenesis of SAKI is so complex that it has not been elucidated. At present, the treatment effect of SAKI is still not ideal, and the mortality rate is high. The main treatment methods include early fluid resuscitation, anti-infection and renal replacement for hypoperfusion of tissues and organs. The earlier the treatment of SAKI is carried out, the higher the survival rate of patients. At present, the diagnosis of SAKI is mainly based on creatinine and urine volume per hour. In the Paper, the purpose of meta-analysis was to analyze the risk factors leading to SAKI and provide help for early identification and prompt treatment of SAKI. We found that the main risk factors for SAKI include diabetes, hypotension (MAP < 65 mmHg), and high serum creatinine (> 88 mol/l). Secondary risk factors include coagulation disorder, liver disease, heart disease, high APACHEII score, lactic acidosis, and oliguria. Our results are generally consistent with other studies, but there are some differences. Such as hypertension, mechanical ventilation, WBC and PCT.

A lot of studies shown that hypotension, high APACHEII score and oliguria are the risk factor of SAKI. (a) Hypotension is a risk factor for SAKI. Because of when patients are in shock, continuous hypotension will lead to severe renal blood insufficiency, which will directly result in renal tubular necrosis, thus causing renal function damage [30]. (b) The assessment items for APACHEII score include age, severe organ dysfunction and immune impairment, GCS score and physiology-related indicators, through which the general situation of patients can better be reflected and the expected mortality of patients can also be evaluated. The higher APACHEII score of patients with sepsis, the worse the basic vital signs and general conditions of patients, and the poorer the body’s self-regulation ability and immune response to infection. According to the research by Cheng Xiaoying et al. [31]. APACHEII score ≥ 19 is an independent risk factor for AKI in patients with sepsis. (c) Oliguria and anuria are also risk factors for the occurrence of SAKI. Systemic hemodynamic instability and reduced renal blood perfusion of patients with sepsis and septic shock lead to renal tissue ischemia and hypoxia and renal tubular damage, resulting in serious reduction of glomerular filtration rate. The decrease of urine volume per hour is a sensitive indicator of renal function damage, which can better predict the occurrence of SAKI.

  1. a.

    Diabetes is a risk factor for SAKI, with unclear mechanism. The possible reasons are that the infection degree of patients with sepsis complicated with diabetes is more serious, and inflammatory factor storm induces the occurrence of AKI. In addition, after severe infection, the incidence of stress hyperglycemia in patients with diabetes increases, and hyperosmolar state induced by high glucose also affects renal function [32].

  2. b.

    Coagulation dysfunction is a risk factor for SAKI. The consumption of coagulation factors and platelets reduces the synthesis of coagulation factors, which promotes the hyperfibrinolysis and leads to the significant extension of PT, APTT and TT and the significant reduction of FIB level, thus resulting indiffuse intravascular coagulation that impairsrenal function. Coagulation dysfunction not only promotes the occurrence of SAKI [33], but also increases the mortality rate of patients with SAKI. Other comorbid factors include heart disease (coronary heart disease and heart failure) and liver disease.

In this study, it is showed that the risk factors of laboratory test for SAKI are lactic acidosis and serum creatinine level. Serum lactic acid level, an important indicator of microcirculation perfusion, can better evaluate the hypoxia state of tissues and organs when it is increased. The increase of creatinine level is a risk factor for SAKI. Creatinine, a macromolecule solute in the body, is a metabolite of creatine, which mainly discharged through glomerular filtration. AKI can be diagnosed when the serum creatinine level of patients with sepsis increases by more than 1.5 times the baseline level, which is of certain value for the early prediction of the occurrence of SAKI [34]. According to the research by Hoste et al. [10], the risk for the development of AKI was increased 7.5 times when serum creatinine level was greater than 1.0 mg/dL (88 mol/L). However, in the process of clinical diagnosis and treatment, the baseline creatinine level of patients is often difficult to be obtained, and the baseline creatinine level at admission does not represent the actual baseline creatinine value. Therefore, the prediction of baseline value of serum creatinine and the degree of elevation in SAKI remains to be studied.

In this study, it is showed that hypertension, invasive mechanical ventilation, WBC and PCT have no statistical significance (P > 0.05). WBC, a traditional infection indicator, is still widely used in the early diagnosis of patients with sepsis, especially bacterial infection. However, there are many factors affecting WBC, and the degree of its elevation may not accurately reflect the severity of sepsis.PCT, an important inflammatory indicator, has higher specificity than other indicators in the early diagnosis of bacterial systemic infection, so it can more accurately assess the severity of the disease. WBC and PCT reflect the severity of sepsis to a certain extent rather than the renal function damage, without predictive value for SAKI. The application of invasive mechanical ventilation should not be used as an indicator reflecting renal function damage. However, the study of Shen Hejin et al. [35] showed that after 24 h of invasive mechanical ventilation, the glomerular filtration rate, urea and creatinine clearance rates of patients decreased significantly. The reason may be that improper use of ventilator, thus resulting in reduced renal blood perfusion and renal failure. At the same time, for patients who use invasive mechanical ventilation for a long time, if the degree of analgesia and sedation is insufficient, the activation of RAAS system can be stimulated by sympathetic nerve excitation and the renal blood flow can be reduced, thus resulting in renal function damage. In the process of clinical treatment, improper use of ventilator rarely occurs. If ventilator parameters are out of the normal range, they will be promptly adjusted.

Although in this study, studies from multiple countries or regions were included, there is no detailed grouping for the distinction between different countries or regions, which may be due to the lack of relevant studies and data on the risk factors for AKI in patients with sepsis in the same country and region. In addition, the research included in this meta-analysis was not completely consistent with the inclusion criteria of the cases, which leads to the heterogeneity of the final results. We have attempted to reduce clinical and methodological heterogeneity by eliminating the weighty literature and conducting a more detailed subgroup analysis of the study, but the analysis effect was not significant and the forest map showed that heterogeneity remained high. In this study, highly heterogeneous data was analyzed by a randomized control model. However, the methods reduce the heterogeneity between studies remain to be studied. In this meta-analysis, publication bias analysis was performed on risk factors (diabetes mellitus, lactic acidosis) that included ≥ 7 articles. The reason of publication bias is that researchers are more likely to publish positive results because negative results have little significance for the study on risk factors. Another possible reason is that some negative studies are not included in databases or journals.

5 Conclusion

The results of this study suggest that the main risk factors for SAKI include diabetes, hypotension (MAP < 65 mmHg), and high serum creatinine (> 88 mol/l). Secondary risk factors include coagulation disorder, liver disease, heart disease, high APACHEII score, lactic acidosis, and oliguria.