Obesity and recovery from acute kidney injury (Ob AKI): a prospective cohort feasibility study

Objectives To test the methodology of recruitment, retention and data completeness in a prospective cohort recruited after a hospitalised episode of acute kidney injury (AKI), to inform a future prospective cohort study examining the effect of obesity on AKI outcomes. Design Feasibility study. Setting Single centre, multi-site UK tertiary hospital. Participants 101 participants (67M; 34F) with a median age of 64 (IQR 53–73) years, with and without obesity, recruited within 3 months of a hospitalised episode of AKI. Outcome measures Feasibility outcomes were recruitment (>15% meeting inclusion criteria recruited), participant retention at 6 and 12 months (≥80%) and completeness of data collection. Exploratory measures included recovery from AKI (regaining >75% of pre-AKI estimated glomerular filtration rate [eGFR]) at 6 months, development or progression of chronic kidney disease (CKD) (kidney function decrease of ≥25% +  rise in CKD category) at 12 months, and associations with poorer kidney outcomes. Results 41% of eligible patients consented to take part, exceeding the target recruitment uptake rate of 15%. Retention was 86% at 6 months and 78% at 12 months; 10 patients died and three commenced dialysis during the study. Data were 90%–100% complete. Median BMI was 27.9 kg/m2 (range 18.1 kg/m2–54.3 kg/m2). 50% of the cohort had stage 3 AKI and 49% had pre-existing CKD. 46% of the cohort met the AKI recovery definition at 6 months. At 12 months, 20/51 patients developed CKD (39%) and CKD progression occurred in 11/49 patients (22%). Post-AKI interleukin-6 and cystatin-C were associated with 12 months decline in eGFR. Conclusions Feasibility to conduct a long-term observational study addressing AKI outcomes associated with obesity was demonstrated. A fully powered prospective cohort study to examine the relationships between obesity and outcomes of AKI is warranted.

individuals there might also be a bias if more severely ill obese vs non obese patients are included with higher fluctuations of weight between admission and discharge. In that regard, it might also be interesting to mention how many of these included patients are ICU patients. It seems appropriate to include patients that are not too heterogeneous and mostly differ in weight rather than in many other confounding variables/ non GFR determinants. -There is a lot of discussion on how to define recovery of kidney function and at what time point this should be done. It might also be difficult to distinguish between AKI still in recovery and CKD. And even in case of apparent recovery there is an increased risk of CKD development. Also, several factors such as severity of AKI, duration of AKI and presence of repetitive episodes of AKI will influence the risk of later CKD development, independent on presence of obesity, and should be taken into account. Presence of proteinuria will also increase the risk of CKD development, even if creatinine levels have returned to baseline. Recovery might be overestimated if there is a large change in body mass which is to be expected in severely ill ICU patients who might demonstrate a decrease in sCr value due to loss of muscle mass rather than to recovery of kidney function -The algorithm used for AKI definition and especially its limitations, should be better elaborated in the text.
-I understand that authors only included patients who have a historical baseline sCr value available, however there is a reason why some patients have this value available and others don't which will inevitably introduce bias. The same can be said for AKI definition with some patients having had more sCr measurements than others during hospitalization. The more values there are, the higher the likelihood of detecting AKI in the first place. -Urinary output is not included in the AKI definition. Although the use of the urinary output criterion in the obese can definitively be problematic, a lot of cases might also be missed by not using this criterion.
-Although the outcome 'change in CKD classification class' is a very practical one, it is rather mathematical and can be artificial e.g. if eGFR changes from 61 to 58 versus from 85 to 46, the change in CKD class will be the same, however these patients have a completely different trajectory.

Reviewer Comment
Response Page Reviewer 1 Examine the possibility of a J shaped relationship between BMI and AKI risk, and consider frailty as a risk factor. Consider further analysis if possible Thank you for your observation that the relationship between BMI and AKI may be J shaped rather than linear. Whilst underweight is likely to be a risk factor for mortality, it is unknown if low BMI is associated with CKD outcomes after AKI. The study is not sufficiently powered for such an analysis.
We acknowledge that frailty may be another risk factor for the development or progression of CKD after an episode of AKI. Frailty could also be considered a risk marker for comorbidity and long-term chronic disease. Our study was a feasibility study with 17 outcomes for recruitment and retention and contained exploratory outcomes related to obesity. There are no measures of frailty in our study and BMI is not a sufficient surrogate.
We have acknowledged frailty as a potential confounder in the discussion. We have suggested that non-linear relationships are considered in future studies.
The methodological comparison of GFR estimation is crucial in this article, I think some improvements will help to understanding easily the difference between creatinine based eGFR using MDRD, CKD-EPI, and cystacin C based eGFR. I suggest the use of abbreviations such as eGFRckd-epi, eGFRcys Thank you for your observation that methodological comparison of GFR estimation is crucial in examining this relationship. We have made clarifications in the text. Only the CKD-EPI equations were used in this study. CKD-EPI equations can be based on creatinine, or cystatin c, or a combination of the two. The MDRD equation was not used in this study. We have adjusted the text and used the following abbreviations: CKD-EPI cystatin c + creatinine equation -eGFRcreat+cysC CKD-EPI creatinine equation -eGFRcreat We did not use the MDRD equation or the CKD-EPI cystatin c based equation in this study.
throughout Reviewer 2 I wonder how authors will be able to differentiate between a 'real' physiological difference in the response of obese vs non obese after an episode of AKI versus differences that are related to misinterpreting serum creatinine values, which might be more likely in the obese vs non obese group.
We agree that interpretation of serum creatinine values with obesity are difficult and have attempted to address this in our study by acknowledging differences in eGFR using different estimating equations. Our examination of the literature indicated that the use of both creatinine and cystatin C in the estimating equation improves the estimation of GFR. We have used the CKD-EPI creatinine + cystatin C equations for all time points after AKI, as cystatin C is not routinely measured.
These problems are exacerbated when weight changes significantly between measurements, particularly when weight decreases, and the error associated with larger body size decreases at the same time.
In our study, body weight increased in all groups over the study period, an expected occurrence after a period of acute illness.
Furthermore, we expect that defining CKD development or progression as at least a 25% drop in eGFR together with a decline in CKD stage, would reflect a "real" physiological difference, rather than only variation in an estimating equation, particularly when intra-individual differences are calculated.
Within individuals there might also be a bias if more severely ill obese vs non obese patients are included with higher fluctuations of weight between admission and discharge. In that regard, it might also be interesting to mention how many of these included patients are ICU patients.
Thank you for raising this important point. Indeed, absolute changes in weight may be higher in obese vs non obese patients who are severely ill. However, typically the percentage weight loss is similar, and this is why definitions of acute malnutrition use percentage weight loss rather than absolute changes in weight to assess risk.
We assessed the severity of the AKI but we did not record whether the admission included an ICU stay or not. We agree that this would be an interesting variable to include in an analysis of the effect of BMI on CKD risk after AKI in a future study.
There is a lot of discussion on how to define recovery of kidney function and at what time point this should be done.
It might also be difficult to distinguish between AKI still in recovery and CKD. And even in case of apparent recovery there is an increased risk of CKD development. Also, several factors such as severity of AKI, duration of AKI and presence of repetitive episodes of AKI will influence the risk of later CKD development, independent on presence of obesity, and should be taken into account.
Thank you for your insightful comments. We agree that the recovery of kidney function is difficult to define and measure. Differences in recovery at 6 months and the development or progression of CKD at 12 months in the present study highlight that recovery or otherwise at 6 months may not be associated with later outcomes.
We completely agree that severity of AKI, duration of AKI, and repetitive AKI episodes will influence the risk of later CKD development. These outcomes were measured in our study, although not reported in this manuscript. We agree that these variables should be taken into account in a study sufficiently large enough to examine risk factors for CKD development and progression after AKI. As our study was a feasibility study on 100 patients, we are unable to adequately examine the impact of these factors on outcomes.
We have acknowledged that these factors should be taken into account in our discussion.

17
Presence of proteinuria will also increase the risk of CKD development, even if creatinine levels have returned to baseline. Recovery might be overestimated if there is a large change in body mass which is to be expected in severely ill ICU patients who might demonstrate a decrease in sCr value due to loss of muscle mass rather than to recovery of kidney function We agree that the serum creatinine may decrease due to a loss of muscle mass after severe illness, and this will impact upon the eGFR. This would be expected in both obese and non-obese patients in severe illness. We have included cystatin-C in our eGFR estimating equation to overcome some of the limitations of using creatinine based equations. As stated above, the short-term recovery from AKI may not necessarily affect longer-term outcomes related to CKD development and progression.
We agree that proteinuria is a likely confounder in the relationship between AKI and CKD. Proteinuria was measured at all timepoints in the current study and it should be included in multivariable analyses in a sufficiently large study. It is beyond the scope of the current study to classify patients by proteinuria status in outcome analyses.
The algorithm used for AKI definition and especially its limitations, should be better elaborated in the text.
We have included a section on the limitations of the KDIGO definition and the use of an algorithm for AKI detection in the discussion.

17
I understand that authors only included patients who have a historical baseline sCr value available, however there is a reason why some patients have this value available and others don't which will inevitably introduce bias. The same can be said for AKI definition with some patients having had more sCr measurements than others during hospitalization. The more values there are, the higher the likelihood of detecting AKI in the first place.
We acknowledge that availability of baseline serum creatinine values and the number of creatinine measurements during hospitalisation does introduce bias. However, this is a limitation of all studies on AKI and is not limited to the present study.

17
Urinary output is not included in the AKI definition. Although the use of the urinary output criterion in the obese can definitively be problematic, a lot of cases might also be missed by not using this criterion.
Urinary output is included in the KDIGO definition of AKI used in the current study.
We have clarified this for the identification of potential participants from referrals to the nephrology service. With the following phase added "using the full KDIGO definition of AKI".
We acknowledge that the urinary criterion is omitted in the AKI detection algorithm used in this study. This is common to many studies on AKI as urinary output is not systematically measured or recorded for all hospitalised patients in a reliable way.

17
Although the outcome 'change in CKD classification class' is a very practical one, it is rather mathematical and can be artificial e.g. if eGFR changes from 61 to 58 versus from 85 to 46, the change in CKD class will be the same, however these patients have a completely different trajectory.
The definition of progression includes both a 25% reduction in eGFR PLUS a decline of at least 1 category in CKD staging within the same patient. Therefore a change in eGFr from 61 to 58 would not meet the full criteria for CKD development or progression. We also looked at the frequency of experiencing a 25% decline in eGFR without a change in CKD category, which could lead to an underestimation of progression, however this occurred in 1 participant only.