This case-control study assesses whether children with type 1 diabetes who experience diabetic ketoacidosis are associated with having a greater risk of developing microalbuminuria and, if so, whether increased risk of microalbuminuria is associated with the occurrence of acute kidney injury during diabetic ketoacidosis.
Key Points
Question
Are episodes of acute kidney injury during diabetic ketoacidosis (DKA) associated with long-term risk of diabetic kidney disease in children with type 1 diabetes?
Findings
In this case-control study of 2345 children with type 1 diabetes, acute kidney injury during DKA was significantly associated with increased risk of development of microalbuminuria. This association persisted after adjusting for age at diagnosis of type 1 diabetes and glycemic control.
Meaning
These data identify a previously unrecognized risk factor for diabetic kidney disease and suggest that even greater efforts should be made to prevent DKA and the occurrence of acute kidney injury during DKA.
Abstract
Importance
Diabetic kidney disease is among the most important causes of end-stage kidney disease worldwide. Risk factors for diabetic kidney disease remain incompletely defined. Recent studies document a high frequency of acute kidney injury (AKI) during diabetic ketoacidosis (DKA) in children, raising the question of whether these AKI episodes might contribute to future risk of diabetic kidney disease.
Objective
To determine whether episodes of AKI occurring during DKA in children are associated with increased risk of development of microalbuminuria.
Design, Setting, and Participants
This retrospective review of medical records included children with type 1 diabetes with 1 or more urine albumin levels measured during routine diabetes care from 2 university-affiliated urban tertiary children’s hospitals in the United States from January 2006 to December 2019. Age at diagnosis of diabetes, hemoglobin A1c levels, episodes of DKA, pH and creatinine levels during DKA, and urine albumin and creatinine measurements were analyzed. Cox proportional hazards regression models were used to identify variables affecting the hazard rate for microalbuminuria development. Analyses began January 2021 and ended May 2021.
Exposures
Episodes of DKA and episodes of AKI occurring during DKA
Main Outcomes and Measures
AKI occurrence and AKI stage were determined from serum creatinine measurements during DKA using Kidney Disease: Improving Global Outcomes criteria. Microalbuminuria was defined as urine albumin–to-creatinine ratio of 30 mg/g or more or excretion of 30 mg or more of albumin in 24 hours.
Results
Of 2345 children, the mean (SD) age at diagnosis was 9.4 (4.4) years. One or more episodes of DKA occurred in 963 children (41%), and AKI occurred during DKA in 560 episodes (47%). In multivariable models adjusting for the associations of age at diagnosis and mean hemoglobin A1c level since diagnosis, each episode of AKI during DKA was associated with a hazard ratio of 1.56 (95% CI, 1.3-1.87) for development of microalbuminuria. Four or more episodes increased the hazard rate by more than 5-fold. DKA episodes without AKI did not significantly increase the hazard rate for microalbuminuria development after adjusting for other covariates.
Conclusions and Relevance
These data demonstrate that episodes of AKI occurring during DKA in children with type 1 diabetes are significantly associated with risk of developing microalbuminuria. Greater efforts are necessary to reduce the frequency of DKA.
Introduction
Diabetic kidney disease (DKD) affects 30% to 40% of adults with diabetes and is the leading cause of end-stage kidney disease in high-resource countries.1 Risk factors for DKD remain incompletely defined despite substantial investigation. Diabetic ketoacidosis (DKA) occurs commonly in children with type 1 diabetes (T1D), and acute kidney injury (AKI) occurs in 43% to 64% of DKA episodes.2,3,4 DKA-related AKI is generally mild, resolves rapidly, and has been attributed to decreased kidney perfusion resulting from intravascular volume depletion.5 Observational studies of children with AKI during episodes of critical illnesses show high frequencies of chronic kidney disease in survivors; however, the severity of AKI in these children is frequently far greater than that observed in DKA.6,7 Whether brief episodes of AKI occurring during DKA contribute to future risk of DKD is unknown.
Microalbuminuria is the earliest manifestation of diabetic nephropathy and a key risk factor for progression to proteinuria in patients with T1D.8 To our knowledge, there are no previous studies investigating the relationship between DKA in children and future development of microalbuminuria. We undertook the current study to determine whether children with T1D who experience DKA have a greater risk of developing microalbuminuria and, if so, whether increased risk of microalbuminuria is associated with the occurrence of AKI during DKA.
Methods
Context and Data Collection
We reviewed the medical records of patients observed at 2 university-affiliated urban academic children’s hospitals. The study protocol was reviewed and approved by the institutional review boards of University of California, Davis and Children’s Hospital of Philadelphia. Because this was a retrospective medical record review, informed consent was not obtained. Children with T1D were identified using databases maintained by the endocrinology divisions at the participating centers and by query of the electronic medical records from January 2006 to December 2019 using International Classification of Diseases, ninth and tenth revisions, for diabetes. We included patients who had T1D diagnosed when younger than 18 years, with at least 1-year duration of diabetes at the time of record review, and having at least 1 urine albumin–to-creatinine ratio (ACR) recorded. We excluded patients with diabetes resulting from causes other than T1D (eg, neonatal diabetes, type 2 diabetes, cystic fibrosis), kidney disease not related to diabetes, and undocumented age at T1D diagnosis. A total of 356 records were excluded for these reasons. If patients transferred care to adult clinics at the study sites, we continued to record data for the duration of follow-up.
We recorded age, sex, age at diagnosis of diabetes, and episodes of DKA (defined as blood glucose concentration >200 mg/dL [to convert to millimoles per liter, multiply by 0.0555] and pH level <7.30 or serum bicarbonate concentration <15 mEq/L [to convert to millimoles per liter, multiply by 1]). We recorded age at the time of each episode and blood pH at presentation. Clinic notes and all outside hospital records accessible through the electronic medical records were reviewed to identify any DKA episodes treated at other hospitals. During each DKA episode, the highest creatinine level measured was used to determine AKI status. As a measure of glycemic control, we calculated the mean of all hemoglobin A1c (HbA1c) levels measured since diagnosis of T1D. For patients with hemoglobinopathies, we recorded fructosamine levels and converted these to equivalent HbA1c levels.9 All urine albumin measurements since diagnosis were recorded (including spot urine ACR and 24-hour measurements). For patients with microalbuminuria, data were recorded only up to the time that microalbuminuria was detected. Data on race and ethnicity were not collected because these were reported inconsistently in the electronic medical record.
Variable and Outcome Definitions
AKI was defined by the Kidney Disease: Improving Global Outcomes serum creatinine criteria.10 Because baseline creatinine values were rarely available, an estimated glomerular filtration rate (GFR) of 120 mL/min/1.73 m2 was used to calculate the expected baseline creatinine level for each patient using the Schwartz estimating equation.11 For these calculations, patients with no recorded height measurements or with clear errors in measured height were assigned height equal to the 50th percentile for the patient’s age and sex (31% of AKI stage calculations). Stages of AKI (0, 1, 2, and 3) were defined by the Kidney Disease: Improving Global Outcomes creatinine cutoffs of 1.5-, 2-, and 3-times estimated baseline creatinine. Moderate/severe DKA was defined by blood pH level less than 7.2 at presentation.
Microalbuminuria was defined as a spot urine ACR of 30 to 299 mg/g.8 Patients who had abnormal spot urine ACRs with subsequently documented elevated ACR on 2 or more spot samples from first morning voids or a 24-hour urine collection (excretion of ≥30 mg of albumin in 24 hours) were considered to have confirmed microalbuminuria. Patients who had abnormal spot urine ACRs with subsequently documented normal urine albumin (2 additional spot urine samples with normal albumin levels or normal albumin levels in 24-hour urine collections) were considered not to have microalbuminuria. Patients with abnormal spot urine ACRs and no follow-up testing were considered to have possible microalbuminuria. These patients were grouped with patients with confirmed microalbuminuria for the main analyses but were removed from the microalbuminuria group in sensitivity analyses to confirm stability of the results. ACRs measured during a hospitalization for DKA were not used to determine microalbuminuria status.
For patients who were not observed at the study sites from the time of diagnosis of T1D, medical records did not always comment on the occurrence of DKA episodes prior to transfer of care. These patients were considered not to have had previous DKA episodes; however, we conducted additional analyses excluding these patients to verify stability of the results.
Statistical Analyses
For each patient, DKA episodes with and without AKI were considered counting processes (ie, 0 from the time of diagnosis until just before the first DKA episode, 1 from the time of the first episode until just before the second episode, etc). In the primary analysis, all AKI stages were counted. In confirmatory analyses, only stages 2 and 3 AKI were considered. We evaluated associations between patient characteristics and the accumulating number of these episodes with time to microalbuminuria using Cox proportional hazards regression models with (when appropriate) time-varying covariates. Time 0 was age at diagnosis of T1D. Patients not developing microalbuminuria were censored at the age of their last ACR measurement. First, we fit models with each predictor variable individually. Then, we fit a multivariable model, including age at diagnosis of T1D, mean HbA1c, DKA episodes without AKI, and DKA episodes with AKI. DKA episodes missing pH measurements were excluded from analyses considering DKA severity, and DKA episodes missing creatinine measurements were excluded from analyses considering AKI. In the primary analysis, we modeled episode counts numerically for simplicity. This approach provides a hazard ratio estimate that corresponds with each additional episode. To determine the association of repeated episodes of DKA with and without AKI, we performed additional analyses with the episode counts considered as categorical variables. For analyses with small numbers of patients having 3 episodes or 4 or more episodes, these categories were combined. All models were tested and residuals inspected for evidence of deviation from the proportional hazards assumption.
In additional analyses that consider confirmed microalbuminuria only, data from patients with possible microalbuminuria (not confirmed) were included in the analysis up to the time of the last normal urine ACR measurement and then censored after that time. Thus, these patients were considered to have normal urine ACRs, but duration of monitoring included only the period up to the last normal measurement.
Finally, the creatinine assay used at 1 site (24% of patient records reviewed) was subject to interference by very high concentrations of acetoacetate. To address this issue, we repeated the analyses with the following modifications to confirm stability of the results: (1) inclusion of study site in the multivariable model, (2) inclusion of an interaction term for site × AKI, and (3) separate analyses for each study site. Two-sided P values were statistically significant at .05. All statistical analyses were performed using R version 3.6.3 (R Foundation). Analyses began January 2021 and ended May 2021.
Results
The study population included 2345 patients with T1D who had received care in the pediatric diabetes clinics at the study sites (eFigure in the Supplement). The mean (SD) age at diagnosis was 9.4 (4.4) years, and the mean (SD) duration of follow-up was 6.5 (3.5) years (Table 1). A total of 963 patients (41%) experienced at least 1 DKA episode, and creatinine measurements were available for 1189 of 1210 episodes (98.3%). AKI occurred in 560 DKA episodes (47%), with stage 1 AKI occurring in 257 AKI events (46%), stage 2 in 238 (43%), and stage 3 in 65 (12%). Overall, 183 patients (7.8%) developed elevated urine ACRs that were not subsequently found to be normal on follow-up testing. Of these 183 patients, microalbuminuria was confirmed by 24-hour urine collection or by multiple samples from first morning voids in 61 individuals (2.6%). The remaining 122 (5.2%) had elevated urine albumin concentrations in single samples but did not have additional testing to either confirm or rule out microalbuminuria. For patients who developed microalbuminuria, the mean duration of diabetes until development of microalbuminuria was 6.5 years (range, 1-23 years).
Table 1. Characteristics of Analytic Cohort (N = 2345)a.
| Characteristics | No. (%) |
|---|---|
| Age at diagnosis of T1D, mean (SD), y | 9.4 (4.4) |
| Follow-up time, mean (SD), y | 6.5 (3.5) |
| Male | 1280 (54.6) |
| Female | 1065 (45.4) |
| HbA1c level since diagnosis of T1D, mean (SD), % of total hemoglobin | 8.3 (1.5) |
| DKA at diagnosis | 799 (34.1) |
| Total DKA episodes experienced | |
| 0 | 1382 (58.9) |
| 1 | 807 (34.4) |
| 2 | 85 (3.6) |
| 3 | 51 (2.2) |
| ≥4 | 20 (0.9) |
| Total moderate or severe DKA episodes experienced | |
| 0 | 1797 (78.0) |
| 1 | 443 (19.2) |
| 2 | 45 (2.0) |
| 3 | 19 (0.8) |
| ≥4 | 2 (0.09) |
| DKA episodes with any degree of AKI | |
| 0 | 1884 (81.0) |
| 1 | 356 (15.3) |
| 2 | 62 (2.7) |
| 3 | 17 (0.7) |
| ≥4 | 6 (0.3) |
| DKA episodes with stage 2 or 3 AKI | |
| 0 | 2079 (89.4) |
| 1 | 206 (8.9) |
| 2 | 28 (1.2) |
| 3 | 10 (0.4) |
| ≥4 | 2 (0.09) |
| Microalbuminuria | |
| No | 2162 (92.2) |
| Possible | 122 (5.2) |
| Confirmed | 61 (2.6) |
Abbreviations: AKI, acute kidney injury; DKA, diabetic ketoacidosis; HbA1c, hemoglobin A1c; T1D, type 1 diabetes.
SI conversion factor: To convert HbA1c to proportion of total hemoglobin, multiply by 0.01.
Data are missing for 40 individuals for DKA severity and missing for 20 individuals for occurrence of AKI during DKA. The mean height for age and sex was used to calculate AKI stage for 372 DKA episodes (31% of DKA episodes analyzed).
The hazard rate for development of microalbuminuria (including both confirmed and possible diagnoses) was higher among patients who were older at the time of diagnosis of T1D and had higher mean HbA1c level (Table 2). The hazard rate for microalbuminuria was also higher among patients after experiencing an episode of DKA and increased further with each subsequent DKA episode (hazard ratio, 1.73; 95% CI, 1.50-2.01). Hazard rate estimates were higher for moderate or severe DKA episodes compared with any DKA episodes. Development of AKI during DKA was also associated with an increased hazard rate for microalbuminuria (hazard ratio for each episode, 1.87; 95% CI, 1.58-2.21) (Figure). Hazard rate estimates for patients with more severe AKI (stage 2 or 3 AKI) were only minimally higher than those for patients with any episodes of AKI.
Table 2. Associations Between Clinical Variables and Risk of Confirmed or Suspected Microalbuminuria.
| Variable | Unadjusted hazard ratio (95% CI)a | P value |
|---|---|---|
| Male (vs female) | 0.88 (0.66-1.17) | .38 |
| Age at diagnosis of diabetes | 1.11 (1.07-1.15) | <.001 |
| Mean HbA1c since diagnosis | 1.37 (1.26-1.49) | <.001 |
| DKA episodes of all severities | ||
| Each episode | 1.73 (1.50-2.01) | <.001 |
| No. of episodes | ||
| 0 | 1 [Reference] | <.001 |
| 1 | 1.38 (1.00-1.92) | |
| 2 | 2.97 (1.67-5.27) | |
| 3 | 7.90 (4.65-13.4) | |
| ≥4 | 4.94 (1.79-13.7) | |
| Moderate or severe DKA | ||
| Each episode | 1.91 (1.58-2.32) | <.001 |
| No. of episodes | ||
| 0 | 1 [Reference] | <.001 |
| 1 | 1.55 (1.10-2.20) | |
| 2 | 3.85 (2.03-7.31) | |
| 3 | 8.67 (4.13-18.2) | |
| ≥4 | 15.6 (2.16-112) | |
| AKI episodes of any severity | ||
| Each episode | 1.87 (1.58-2.21) | <.001 |
| No. of episodes | ||
| 0 | 1 [Reference] | <.001 |
| 1 | 1.33 (0.92-1.94)b | |
| 2 | 4.52 (2.65-7.73) | |
| 3 | 8.20 (4.03-16.7) | |
| ≥4 | 9.31 (2.87-30.2) | |
| Stage 2 or 3 AKI | ||
| Each episode | 1.98 (1.62-2.42) | <.001 |
| No. of episodes | ||
| 0 | 1 [Reference] | <.001 |
| 1 | 1.54 (1.02-2.33) | |
| 2 | 4.82 (2.50-9.30) | |
| 3 | 8.67 (3.85-19.5) | |
| ≥4 | 15.0 (2.08-108) |
Abbreviations: AKI, acute kidney injury; DKA, diabetic ketoacidosis; HbA1c, hemoglobin A1c.
Hazard ratio represents increased risk for each 1-year increase in age at diagnosis and each 1% increase in mean HbA1c level since diagnosis.
Hazard ratio values reported represent estimates when patients have experienced exactly 1, 2, 3 or 4 episodes or more of acute kidney injury, compared with 0 episodes. Sample sizes at some levels were small, affecting power to detect significant differences for individual levels; however, the overall comparison across levels indicated a significant association of number of acute kidney injury episodes with risk of microalbuminuria.
Figure. Microalbuminuria Development in Patients With Type 1 Diabetes With and Without History of AKI During Diabetic Ketoacidosis.
This graph depicts the proportion of patients remaining without microalbuminuria after diagnosis of type 1 diabetes (n = 2325). Twenty patients were excluded owing to missing data regarding occurrence of acute kidney injury (AKI) during diabetic ketoacidosis episodes.
In multivariable models adjusting for age at diagnosis of T1D and mean HbA1c, DKA with AKI continued to be significantly associated with an increased hazard rate for microalbuminuria (hazard ratio for each episode, 1.56; 95% CI, 1.30-1.87). However, DKA without AKI was not significantly associated with the hazard rate for microalbuminuria (hazard ratio for each episode, 1.22; 95% CI, 0.94-1.57) (Table 3). Hazard rates for development of microalbuminuria in the multivariable models were slightly lower than those in univariable analyses, likely reflecting the predictive value of age and HbA1c level that are independently associated with microalbuminuria. In a separate multivariable model, where each episode number was considered a category, the hazard rate increased by more than 3-fold for patients with 2 AKI episodes and more than 5-fold for patients with 4 or more AKI episodes (Table 3).
Table 3. Multivariable Models of Factors Associated With Microalbuminuria.
| Variable | Adjusted hazard ratio (95% CI)a | P valueb |
|---|---|---|
| Multivariable model: risk of confirmed or suspected microalbuminuria | ||
| Age at diagnosis of T1D | 1.10 (1.06-1.15) | <.001 |
| Mean HbA1c level since diagnosis of T1D | 1.27 (1.16-1.39) | <.001 |
| DKA with no AKI | 1.22 (0.94-1.57) | .14 |
| DKA with AKI | 1.56 (1.30-1.87) | <.001 |
| Multivariable model with DKA and AKI variables analyzed as categories based on frequency | ||
| Age at diagnosis of T1D | 1.11 (1.07-1.15) | <.001 |
| Mean HbA1c level since diagnosis of T1D | 1.27 (1.16-1.39) | <.001 |
| DKA episode with no AKI | ||
| 0 | 1 [Reference] | .50 |
| 1 | 1.34 (0.91-1.98) | |
| 2 | 1.30 (0.56-3.03) | |
| ≥3c | 0.91 (0.12-6.70) | |
| DKA episode with AKI | ||
| 0 | 1 [Reference] | <.001 |
| 1 | 1.29 (0.88-1.90) | |
| 2 | 3.12 (1.77-5.50) | |
| 3 | 3.73 (1.72-8.11) | |
| ≥4 | 5.47 (1.67-17.9) |
Abbreviations: AKI, acute kidney injury; DKA, diabetic ketoacidosis; HbA1c, hemoglobin A1c; T1D, type 1 diabetes.
Hazard ratio represents increased risk for each 1-year increase in age at diagnosis and each 1% increase in mean HbA1c level since diagnosis.
P value from a likelihood ratio test with other variables in the model.
Because the number of patients with 4 or more episodes of DKA without AKI was low, causing statistical instability in the multivariable model, this category was combined with 3 episodes.
In a sensitivity analysis that censored patients with possible microalbuminuria (microalbuminuria measured in a single sample that was not subsequently confirmed with additional testing), we repeated the univariable (eTable in the Supplement) and multivariable analyses (Table 4). These results were similar to the main analyses with AKI episodes during DKA significantly increasing the hazard rate for microalbuminuria.
Table 4. Multivariable Model: Risk of Confirmed Microalbuminuria (n = 2333).
| Variable | Adjusted HR (95% CI)a | P value |
|---|---|---|
| Age at diagnosis of T1D | 1.03 (0.96-1.11) | .38 |
| Mean HbA1c level since diagnosis of T1D | 1.46 (1.25-1.69) | <.001 |
| DKA with no AKI | 0.89 (0.52-1.54) | .68 |
| DKA with AKI | 1.66 (1.25-2.20) | <.001 |
Abbreviations: AKI, acute kidney injury; DKA, diabetic ketoacidosis; HbA1c, hemoglobin A1c; HR, hazard ratio; T1D, type 1 diabetes.
Hazard ratio represents increased risk for each 1-year increase in age at diagnosis and each 1% increase in mean HbA1c level since diagnosis.
Finally, to account for differences in creatinine assays used at the study sites and to ensure that possible interference of high acetoacetate levels with one of the assays did not influence the results, we conducted several additional analyses including site in the multivariable model, including an interaction term involving site and AKI, or analyzing data from each study site separately. In all subanalyses, the results were unchanged, with AKI, age at T1D diagnosis, and mean HbA1c level maintaining significant associations with risk of microalbuminuria (data not shown).
Discussion
Diabetes is an important cause of end-stage kidney disease worldwide, but risk factors for kidney dysfunction resulting from T1D remain incompletely defined. Recent studies document a high frequency of AKI in children during DKA.2,3,4 In the current study, we found that the occurrence of AKI during DKA was associated with an increased risk for development of microalbuminuria. After adjusting for glycemic control and age at diagnosis of T1D in multivariable analyses, a single episode of AKI increased the hazard rate for microalbuminuria by 56%. Multiple episodes increased the hazard rate by as much as 5-fold. Episodes of DKA without occurrence of AKI were not significantly associated with the risk of microalbuminuria. To our knowledge, these data identify a previously unrecognized risk factor for DKD.
Risk factors for DKD include hyperglycemia,12,13 insulin resistance,14 hypertension,15 and elevated uric acid levels.16 Microalbuminuria is the earliest marker of DKD and occurs in 26% of pediatric patients after 10 years’ duration of diabetes and in 51% after 19 years.17 Previous studies in children and adolescents with T1D have documented increased risk of microalbuminuria among patients with higher HbA1c level and longer duration of T1D, similar to the findings of the current study.12,13 Higher blood pressure, dyslipidemia, and high normal urine albumin excretion have also been associated with risk of microalbuminuria in children and adolescents.13,18
A few studies in adults with diabetes have identified connections between AKI and progression of chronic kidney disease (CKD).19,20 In a retrospective study, AKI occurring during a hospitalization was associated with increased risk of developing advanced CKD.19 However, the main reason for hospitalization in this cohort was cardiovascular disease, and it was not noted whether any patients were hospitalized for DKA. In a retrospective study of 179 adults with DKA, the occurrence of AKI during DKA was associated with rapid progression of CKD and with long-term mortality.20 Patients with preexisting CKD (11% of the sample) were more likely to develop AKI during DKA. Although the presence of CKD was determined from review of medical records, the presence of microalbuminuria prior to hospitalization was not documented. Therefore, it was unclear whether the occurrence of AKI reflected greater susceptibility to AKI in patients with underlying DKD or whether kidney damage related to AKI contributed to disease progression. In the current study, data collection began at the time of diagnosis of T1D and ended when microalbuminuria developed, such that all recorded episodes of DKA and AKI preceded development of microalbuminuria. These data strengthen the evidence for a causative association between DKA-related AKI and microalbuminuria.
In a 2020 study of children with DKA, development of AKI during DKA was associated with increased risk of AKI during subsequent DKA episodes by 9-fold.3 These data suggest that episodes of AKI may not only increase risk of CKD but may also make the kidneys more susceptible to future AKI. In this same study, occurrence of AKI during DKA was also associated with lower IQ scores after recovery, raising the possibility that DKA may be associated with a syndrome of multiple organ dysfunction affecting both brain and kidneys and raising the question of whether similar processes might contribute to declines in both kidney and cognitive function in patients with T1D.3
Because AKI episodes also contribute to risk of albuminuria and CKD in conditions other than diabetes, the current study raises the question of whether AKI during DKA increases risk of DKD or causes kidney injury via a separate mechanism with similar end points. The major mechanism driving the transition from AKI to CKD in other conditions involves development of tubulointerstitial fibrosis following proximal tubular injury.21,22 Glomerular endothelial dysfunction, oxidative stress, and ongoing inflammation are also implicated.23,24 DKA causes a hyperinflammatory state with elevations of several cytokines that have also been implicated in the pathogenesis of DKD.25,26,27,28,29,30 Interestingly, animal data suggest that DKA might trigger a chronic neuroinflammatory response31; however, persistent kidney inflammation after DKA has not been investigated.
Limitations
The current study has several limitations inherent to retrospective studies. ACR measurements were not standardized such that testing began at varying times after diagnosis of T1D and occurred at varying intervals. Therefore, it is possible that detection of microalbuminuria was delayed in some cases. Transient microalbuminuria that resolved without intervention may also have been missed but these episodes are of unclear clinical relevance. Methods for collecting urine ACR samples also varied and samples from the first morning void may have generated more accurate results than randomly collected samples. However, results from subanalyses involving only patients with confirmed microalbuminuria were similar to the main analyses. In addition, baseline creatinine levels were estimated, and the ideal GFR assumption to determine baseline creatinine is not uniformly agreed on. We used an assumed GFR of 120 mL/min/1.73 m2 to be consistent with other pediatric AKI studies.2,32,33 Use of lower estimates for GFR would have resulted in fewer patients being categorized as having AKI. However, the results of our analyses were similar when only stage 2 or 3 AKI episodes were included in the AKI group. In addition, we used the 50th percentile for age and sex as an estimate of height to calculate AKI stage in patients with missing height data. This may have over- or underestimated AKI severity in some patients. In addition, some DKA episodes treated at other hospitals may not have been recorded and therefore may have been missed. The timing and frequency of creatinine measurements during DKA also varied such that some AKI episodes may have been missed or the severity underestimated. However, it is unlikely that data on any covariates were systematically missing from specific population subgroups, such that missing data would have tended to obscure or diminish the estimated association of AKI episodes with microalbuminuria, rather than producing spurious associations. More precise estimates of the association of AKI episodes with risk of microalbuminuria could be obtained from a prospective study with standardized monitoring methods.
Although microalbuminuria is used to identify patients at risk of CKD,8 microalbuminuria may be reversible and does not necessarily predict permanent kidney damage.34,35 Furthermore, microalbuminuria is not a universal finding in patients who develop DKD.36,37,38 Therefore, use of microalbuminuria as the only indicator of DKD may falsely identify or fail to identify some patients with DKD.39,40,41 Nonetheless, measurement of urinary albumin excretion remains the standard of care for evaluating potential kidney injury in children with diabetes, and children with microalbuminuria would be expected to have a higher frequency of DKD than those without microalbuminuria.35
Conclusions
The occurrence of AKI during DKA was associated with a substantially increased hazard rate for development of microalbuminuria. Multiple episodes of AKI increased the hazard rate as much as 5-fold. These data suggest that AKI during DKA contributes to the development of DKD and underscore the importance of DKA prevention in improving health outcomes for children with diabetes.
eFigure. Frequency of AKI and microalbuminuria among patients grouped according to number of DKA episodes
eTable. Associations between clinical variables and risk of confirmed microalbuminuria
References
- 1.Alicic RZ, Rooney MT, Tuttle KR. Diabetic kidney disease: challenges, progress, and possibilities. Clin J Am Soc Nephrol. 2017;12(12):2032-2045. doi: 10.2215/CJN.11491116 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hursh BE, Ronsley R, Islam N, Mammen C, Panagiotopoulos C. Acute kidney injury in children with type 1 diabetes hospitalized for diabetic ketoacidosis. JAMA Pediatr. 2017;171(5):e170020. doi: 10.1001/jamapediatrics.2017.0020 [DOI] [PubMed] [Google Scholar]
- 3.Myers SR, Glaser NS, Trainor JL, et al. ; Pediatric Emergency Care Applied Research Network (PECARN) DKA FLUID Study Group . Frequency and risk factors of acute kidney injury during diabetic ketoacidosis in children and association with neurocognitive outcomes. JAMA Netw Open. 2020;3(12):e2025481. doi: 10.1001/jamanetworkopen.2020.25481 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Huang SKHC, Huang CY, Lin CH, et al. Acute kidney injury is a common complication in children and adolescents hospitalized for diabetic ketoacidosis. PLoS One. 2020;15(10):e0239160. doi: 10.1371/journal.pone.0239160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Orban JCME, Maizière EM, Ghaddab A, Van Obberghen E, Ichai C. Incidence and characteristics of acute kidney injury in severe diabetic ketoacidosis. PLoS One. 2014;9(10):e110925. doi: 10.1371/journal.pone.0110925 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Greenberg JH, Coca S, Parikh CR. Long-term risk of chronic kidney disease and mortality in children after acute kidney injury: a systematic review. BMC Nephrol. 2014;15:184. doi: 10.1186/1471-2369-15-184 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Mammen C, Al Abbas A, Skippen P, et al. Long-term risk of CKD in children surviving episodes of acute kidney injury in the intensive care unit: a prospective cohort study. Am J Kidney Dis. 2012;59(4):523-530. doi: 10.1053/j.ajkd.2011.10.048 [DOI] [PubMed] [Google Scholar]
- 8.Donaghue KCMM, Marcovecchio ML, Wadwa RP, et al. ISPAD Clinical Practice Consensus Guidelines 2018: microvascular and macrovascular complications in children and adolescents. Pediatr Diabetes. 2018;19(suppl 27):262-274. doi: 10.1111/pedi.12742 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Guillausseau P-JCM-A, Charles MA, Godard V, et al. Comparison of fructosamine with glycated hemoglobin as an index of glycemic control in diabetic patients. Diabetes Res. 1990;13(3):127-131. [PubMed] [Google Scholar]
- 10.Acute Kidney Injury Work Group . KDIGO clinical practice guideline for acute kidney injury. Kidney Int Suppl. 2012;2:1-138. [Google Scholar]
- 11.Schwartz GJ, Muñoz A, Schneider MF, et al. New equations to estimate GFR in children with CKD. J Am Soc Nephrol. 2009;20(3):629-637. doi: 10.1681/ASN.2008030287 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Diabetes Control and Complications Trial Research Group . Effect of intensive diabetes treatment on the development and progression of long-term complications in adolescents with insulin-dependent diabetes mellitus: Diabetes Control and Complications Trial. J Pediatr. 1994;125(2):177-188. doi: 10.1016/S0022-3476(94)70190-3 [DOI] [PubMed] [Google Scholar]
- 13.Daniels M, DuBose SN, Maahs DM, et al. ; T1D Exchange Clinic Network . Factors associated with microalbuminuria in 7,549 children and adolescents with type 1 diabetes in the T1D Exchange clinic registry. Diabetes Care. 2013;36(9):2639-2645. doi: 10.2337/dc12-2192 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Helliwell R, Warnes H, Kietsiriroje N, et al. Body mass index, estimated glucose disposal rate and vascular complications in type 1 diabetes: Beyond glycated haemoglobin. Diabet Med. 2021;38(5):e14529. doi: 10.1111/dme.14529 [DOI] [PubMed] [Google Scholar]
- 15.UK Prospective Diabetes Study Group . Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38: UK Prospective Diabetes Study Group. BMJ. 1998;317(7160):703-713. [PMC free article] [PubMed] [Google Scholar]
- 16.Bjornstad P, Laffel L, Lynch J, et al. ; TODAY Study Group . Elevated serum uric acid is associated with greater risk for hypertension and diabetic kidney diseases in obese adolescents with type 2 diabetes: an observational analysis from the Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY) study. Diabetes Care. 2019;42(6):1120-1128. doi: 10.2337/dc18-2147 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Amin R, Widmer B, Prevost AT, et al. Risk of microalbuminuria and progression to macroalbuminuria in a cohort with childhood onset type 1 diabetes: prospective observational study. BMJ. 2008;336(7646):697-701. doi: 10.1136/bmj.39478.378241.BE [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Stone MLCM, Craig ME, Chan AK, Lee JW, Verge CF, Donaghue KC. Natural history and risk factors for microalbuminuria in adolescents with type 1 diabetes: a longitudinal study. Diabetes Care. 2006;29(9):2072-2077. doi: 10.2337/dc06-0239 [DOI] [PubMed] [Google Scholar]
- 19.Thakar CV, Christianson A, Himmelfarb J, Leonard AC. Acute kidney injury episodes and chronic kidney disease risk in diabetes mellitus. Clin J Am Soc Nephrol. 2011;6(11):2567-2572. doi: 10.2215/CJN.01120211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chen J, Zeng H, Ouyang X, et al. The incidence, risk factors, and long-term outcomes of acute kidney injury in hospitalized diabetic ketoacidosis patients. BMC Nephrol. 2020;21(1):48. doi: 10.1186/s12882-020-1709-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Venkatachalam MAWJ, Weinberg JM, Kriz W, Bidani AK. Failed tubule recovery, AKI-CKD transition, and kidney disease progression. J Am Soc Nephrol. 2015;26(8):1765-1776. doi: 10.1681/ASN.2015010006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Nath KA. Tubulointerstitial changes as a major determinant in the progression of renal damage. Am J Kidney Dis. 1992;20(1):1-17. doi: 10.1016/S0272-6386(12)80312-X [DOI] [PubMed] [Google Scholar]
- 23.Fiorentino M, Grandaliano G, Gesualdo L, Castellano G. Acute kidney injury to chronic kidney disease transition. Contrib Nephrol. 2018;193:45-54. doi: 10.1159/000484962 [DOI] [PubMed] [Google Scholar]
- 24.Yu SMBJ, Bonventre JV. Acute kidney injury and progression of diabetic kidney disease. Adv Chronic Kidney Dis. 2018;25(2):166-180. doi: 10.1053/j.ackd.2017.12.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Stentz FB, Umpierrez GE, Cuervo R, Kitabchi AE. Proinflammatory cytokines, markers of cardiovascular risks, oxidative stress, and lipid peroxidation in patients with hyperglycemic crises. Diabetes. 2004;53(8):2079-2086. doi: 10.2337/diabetes.53.8.2079 [DOI] [PubMed] [Google Scholar]
- 26.Li J, Huang M, Shen X. The association of oxidative stress and pro-inflammatory cytokines in diabetic patients with hyperglycemic crisis. J Diabetes Complications. 2014;28(5):662-666. doi: 10.1016/j.jdiacomp.2014.06.008 [DOI] [PubMed] [Google Scholar]
- 27.Ozer G, Teker Z, Cetiner S, et al. Serum IL-1, IL-2, TNFalpha and INFgamma levels of patients with type 1 diabetes mellitus and their siblings. J Pediatr Endocrinol Metab. 2003;16(2):203-210. doi: 10.1515/JPEM.2003.16.2.203 [DOI] [PubMed] [Google Scholar]
- 28.Omatsu T, Cepinskas G, Clarson C, et al. ; Canadian Critical Care Translational Biology Group . CXCL1/CXCL8 (GROα/IL-8) in human diabetic ketoacidosis plasma facilitates leukocyte recruitment to cerebrovascular endothelium in vitro. Am J Physiol Endocrinol Metab. 2014;306(9):E1077-E1084. doi: 10.1152/ajpendo.00659.2013 [DOI] [PubMed] [Google Scholar]
- 29.Karavanaki K, Karanika E, Georga S, et al. Cytokine response to diabetic ketoacidosis (DKA) in children with type 1 diabetes (T1DM). Endocr J. 2011;58(12):1045-1053. doi: 10.1507/endocrj.EJ11-0024 [DOI] [PubMed] [Google Scholar]
- 30.Donate-Correa J, Ferri CM, Sánchez-Quintana F, et al. Inflammatory cytokines in diabetic kidney disease: pathophysiologic and therapeutic implications. Front Med (Lausanne). 2021;7(7):628289. doi: 10.3389/fmed.2020.628289 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Glaser N, Chu S, Hung B, et al. Acute and chronic neuroinflammation is triggered by diabetic ketoacidosis in a rat model. BMJ Open Diabetes Res Care. 2020;8(2):e001793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Basu RK, Kaddourah A, Terrell T, et al. ; Prospective Pediatric AKI Research Group (ppAKI) . Assessment of Worldwide Acute Kidney Injury, Renal Angina and Epidemiology in critically ill children (AWARE): study protocol for a prospective observational study. BMC Nephrol. 2015;16:24. doi: 10.1186/s12882-015-0016-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Zappitelli M, Parikh CR, Akcan-Arikan A, Washburn KK, Moffett BS, Goldstein SL. Ascertainment and epidemiology of acute kidney injury varies with definition interpretation. Clin J Am Soc Nephrol. 2008;3(4):948-954. doi: 10.2215/CJN.05431207 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Perkins BAFL, Ficociello LH, Silva KH, Finkelstein DM, Warram JH, Krolewski AS. Regression of microalbuminuria in type 1 diabetes. N Engl J Med. 2003;348(23):2285-2293. doi: 10.1056/NEJMoa021835 [DOI] [PubMed] [Google Scholar]
- 35.de Boer IHRT, Rue TC, Cleary PA, et al. ; Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Study Research Group . Long-term renal outcomes of patients with type 1 diabetes mellitus and microalbuminuria: an analysis of the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications cohort. Arch Intern Med. 2011;171(5):412-420. doi: 10.1001/archinternmed.2011.16 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Tsalamandris C, Allen TJ, Gilbert RE, et al. Progressive decline in renal function in diabetic patients with and without albuminuria. Diabetes. 1994;43(5):649-655. doi: 10.2337/diab.43.5.649 [DOI] [PubMed] [Google Scholar]
- 37.Retnakaran R, Cull CA, Thorne KI, Adler AI, Holman RR; UKPDS Study Group . Risk factors for renal dysfunction in type 2 diabetes: U.K. Prospective Diabetes Study 74. Diabetes. 2006;55(6):1832-1839. doi: 10.2337/db05-1620 [DOI] [PubMed] [Google Scholar]
- 38.Perkins BAFL, Ficociello LH, Roshan B, Warram JH, Krolewski AS. In patients with type 1 diabetes and new-onset microalbuminuria the development of advanced chronic kidney disease may not require progression to proteinuria. Kidney Int. 2010;77(1):57-64. doi: 10.1038/ki.2009.399 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Ye X, Luo T, Wang K, et al. Circulating TNF receptors 1 and 2 predict progression of diabetic kidney disease: a meta-analysis. Diabetes Metab Res Rev. 2019;35(8):e3195. doi: 10.1002/dmrr.3195 [DOI] [PubMed] [Google Scholar]
- 40.Papadopoulou-Marketou N, Margeli A, Papassotiriou I, Chrousos GP, Kanaka-Gantenbein C, Wahlberg J. NGAL as an early predictive marker of diabetic nephropathy in children and young adults with type 1 diabetes mellitus. J Diabetes Res. 2017;2017:7526919. doi: 10.1155/2017/7526919 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Tang XYZJ, Zhou JB, Luo FQ, et al. Urine NGAL as an early biomarker for diabetic kidney disease: accumulated evidence from observational studies. Ren Fail. 2019;41(1):446-454. doi: 10.1080/0886022X.2019.1617736 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eFigure. Frequency of AKI and microalbuminuria among patients grouped according to number of DKA episodes
eTable. Associations between clinical variables and risk of confirmed microalbuminuria