Mortality and renal prognosis in isolated metformin‐associated lactic acidosis treated with continuous renal replacement therapy and citrate‐calcium‐anticoagulation

Use of metformin increases plasma lactate concentration and may lead to metformin‐associated lactic acidosis (MALA). Previous studies have suggested severe MALA to have a mortality of 17%‐21%, but have included patients with other coincident conditions such as sepsis. The treatment of choice is continuous renal replacement therapy (CRRT), which has been performed using heparin analogues or no anticoagulation in former studies.

Patients with B-type lactic acidosis likely due to isolated MALA were screened by carefully excluding other potential causes of lactic acidosis. Therefore, patients with septic or cardiogenic shock, arterial thrombosis leading to end-tissue hypoxemia, significant rhabdomyolysis, other medications, or substances such as acetaminophen, ethylene glycol, or methanol potentially responsible for lactic acidosis and patients with hypermetabolic lactic acidosis due to hematological malignancies were excluded from the main analysis. None of the patients considered to have isolated MALA had been diagnosed with or suspected to have sepsis and none had positive blood cultures. A total of 45 patients with metformin and type B lactic acidosis during the study period were screened for inclusion and 23 patients were considered to have isolated MALA.
For the purpose of this study, blood pH, bicarbonate, lactate, electrolytes, glucose and blood pressure, vasoactive medication and diuresis, were recorded at 6-h intervals from the ICU admission to 72 hours following admission in the isolated MALA group.
Other data extracted from patients' medical records included reason for ICU admission, demographics, medical conditions (both chronic and acute), medications, and other relevant laboratory results. Creatinine and eGFR were assessed at baseline (within 1 year prior to ICU admission), on admission and at hospital discharge and at 1 year of follow-up in survivors (available follow-up creatinine closest to one year, but at least 6 months following MALA).

| CRRT modality
CVVHD for all patients was performed according to a standard protocol employed in our centre using Fresenius Multifiltrate CRRT monitors and 1.80 m2 polysulfone hemofilter Ultraflux EMiC2 HCO membranes with CiCa dialysate to achieve regional citrate anticoagulation (Fresenius Medical Care, Bad Hamburg, Germany).
Post-filter-ionized calcium levels were used for anticoagulation monitoring. The starting citrate dose used was 4 mmol per liter of treated blood-a standard dose recommended for CiCa-CVVHD by the manufacturer. The citrate dose was then titrated during the treatment to achieve a target Ca++ level and thereby sufficient regional anticoagulation without systemic effects. Blood and dialysate flow rates were prescribed according to the weight of the patient and by the caring ICU physician to target a dialysis dose of 30 mL/kg/h.

Editorial Comment
The authors report from a single center how 23 patients with isolated metformin-associated lactic acidosis were successfully treated with continuous veno-venous hemodialysis using regional citrate-calcium-anticoagulation. In 1-year follow-up, survival was good, but approximately a third developed chronic kidney disease.

| Statistical methods
Results are expressed as mean ± standard deviation (SD) for normally distributed variables and median [interquartile range (IQR)] for skewed continuous variables, unless stated otherwise. Univariate associations between the study variables were analyzed by calculating the Pearson's correlation coefficients. Comparisons between creatinine and eGFR, respectively, at baseline and 1 year following discharge were done using Wilcoxon signed-rank tests. The between group comparisons concerning ICU, 90-day, and 1-year mortality were performed using Fisher's exact test. All statistical analyses were performed using statistical analysis system, SAS version 9.3 (SAS Institute Inc, Cary NC).

| Compliance with ethical standards
The study was approved by the Hospital District of Southwest Finland (Reference number: T143/2016). This was a retrospective registerbased study of patients from an anonymized dataset that only involved recording data from medical records. According to Finnish law and Ethics committee of South-West Finland Hospital District this study did not require consent from patients to participate.

| Data availability
Data that support the findings of this study are available from the  In the remaining 22 patients, another acute comorbidity (apart from metformin-use) was considered as the primary etiology for lactic acidosis. Etiologies for lactic acidosis of the remaining patients included sepsis, cardiogenic shock, liver failure, brain stem infarction, and severe ethanol poisoning.
The patients that were excluded from the analyses due to acute comorbidities had significantly higher ICU (31.8%, P = .004), 90-day   19 The low mortality observed in the current study is most likely explained by exclusion of patients with sepsis, septic shock, liver failure, cardiac failure, and marked chronic renal disease on contrary to previous studies. It is plausible that in a notable proportion of patients included in previous studies, metformin has been a bystander or only partly responsible for the lactic acidosis observed in the critically ill patients with other acute comorbidities. In line with this assumption, the patient group with other acute comorbities (sepsis, cardiovascular, hepatic etc) had a markedly higher ICU, 90day and 1-year mortality in the present study compared to patients with isolated MALA. In these comorbid patients mortality was similar to earlier reports 15,16 The safe use of metformin in patients with other chronic risk factors for developing lactic acidosis is still under debate.

| D ISCUSS I ON
Available data suggest cautious use of metformin in patients with mild to moderate kidney disease, 8  was also associated with the duration of ICU stay. Blood lactate, however, declined rapidly using CRRT with regional citrate anticoagulation with two-thirds of patients achieving normal lactate and pH by 24 hours of ICU admission. Peak serum creatinine level has been earlier shown to be associated with mortality of MALA patients. 15 The mortality in our study was so rare that its determinants could not be studied. Instead creatinine at discharge was directly correlated with duration of ICU stay, ventilator treatment, and duration RRT, and inversely with pH at 24 hours of admittance to the ICU. The novel finding of the current study was that eGFR and creatinine re- Our study has limitations. We were not able to measure metformin concentrations, since these are not routinely measured in Finland. It could have been useful to study the association between metformin concentrations and the severity of MALA and to know whether toxic concentrations of metformin were present. However, previous studies have not associated metformin dose or plasma concentrations with prognosis of MALA. 15 Furthermore, many of the previous studies have included several chronic and/or acute conditions potentially confounding the primary cause of lactic acidosis, a scenario that we aimed to carefully exclude. Therefore, we believe we were able to screen all isolated MALA patients without confounding conditions during a 10-year period treated with CRRT and citrate-calcium-anticoagulation in our hospital district involving 470 000 citizens.
Our current results clearly show that the prognosis of patients with isolated MALA is good although renal function may remain impaired from baseline levels after an episode of MALA.

ACK N OWLED G EM ENTS
The authors are grateful to Mrs Eveliina Loikas, RN for help with the data collection.

CO N FLI C T O F I NTE R E S T S
P. Uusalo has received honoraria for speaking at symposia organized by Orion Pharma. M. Järvisalo declares no conflicts of interest.