A cross-sectional study was conducted with 100 patients admitted in KMC group of hospitals between July 2014 and August 2015. Written informed consent was taken from all patients included in the study. All patients more than 18 years of age and diagnosed of AKI as per KDIGO criteria were included [6]. Patients with multiorgan dysfunction and those who were on the ventilator, postrenal (obstructive uropathy), and toxin induced (drug), hypokalemia, hypocalcaemia and diabetics were excluded. AKI is defined as increase in serum creatinine level x 3.0, decrease in GFR by 75%, or serum creatinine level ≥4 mg/dL with acute increase of >0.5 mg/dL; UO <0.3 mL/kg/hour for 24 hours, or anuria for 12 hours. Patients whose creatinine was normalised at the end of follow up of study (<1.2 mg/dL) were considered under the recovery. Patients whose creatinine level was not normalised at the end of follow up were considered to be non recovered. Serum magnesium levels were measured on day of admission, day 3 and day 6 and clinical features of hypomagnesaemia were observed. Magnesium <1.5 mg/dL was considered as hypomagnesaemia. Magnesium levels were measured using colorimetric method by COBAS 601/501 analyser.

Data were analysed by Statistical Package of Social Science (SPSS) version 17.0 and the mean and proportion values of various variables were calculated accordingly using Chi-square test. A p-value less than 0.05 were considered as statistically significant.

In total 140 cases, 40 cases were excluded out of which 10 were due to obstructive uropathy, 10 were on the ventilator, 10 were suffering from multiorgan dysfunction syndrome. Ten patients were treated with 3 sessions of dialysis. Majority of the patients were males (n=73, 73.0%). Most of the patients belonged to the age group 41 years and above (n=54, 54.0%) [Table/Fig-1]. Symptomatic treatment was given in 100 patients and they recovered. No specific drugs were given including nephrotoxic drugs. Prevalence of hypomagnesaemia was 69%, 43%, and 27% on day 1, day 3, and day 6 respectively. Out of n=69 patients with hypomagnesaemia on day 1, n=60 patients recovered. Hypomagnesaemia at day 1 was significantly (p=0.004) associated with recovery from AKI. Out of 43 patients with hypomagnesaemia on day 3, 38 recovered and 5 did not recover. These observations were significant (p=0.001). On Day 6, out of 27 patients with hypomagnesaemia, 23 recovered and 4 were non recovered (p=0.001). In patients who have recovered, day 1, mean magnesium levels in recovery and non recovery were 1.29 and 1.79 meq/L respectively which is highly significant (p=0.001) [Table/Fig-2]. Out of 31 patients with serum magnesium levels >1.5 meq/L, on day 1, 12 were non recovered (38.7%, p=0.004) showing an association with non recovery of AKI. Serum magnesium levels >1.5 meq/L on day 3 was not significantly (p=0.133) associated with non recovery of AKI. Out of 57 patients with serum magnesium levels >1.5 meq/L, 15 were non recovered and 42 were recovered (26.3% vs. 73.7%). Among patients with serum magnesium levels >1.5 meq/L on D6, out of 73, 17 were non recovered and 56 were recovered (23.3% vs. 76.7%, p=0.356). In patients who recovered, mean day 1 serum creatinine was 4.1 mg/dL with a standard deviation of 1.62. In patients who have non recovered, mean day 1 serum creatinine was 6.2 with a standard deviation of 2.4. These values are significant (p=0.001) [Table/Fig-3].

Though the present study aimed to assess the prevalence of hypomagnesaemia in AKI and association of serum magnesium levels in non recovery of AKI. We found that the prevalence of hypomagnesaemia was 69% on day 1, 43% on day 3, 27% on day 6 respectively. Hypomagnesaemia at day 1 in this study was significantly associated with recovery (n=60, p=0.004). Hypomagnesaemia is a frequent electrolyte abnormality found in inpatients in a hospital, especially in the seriously ill patients. Hypomagnesaemia has high risk of cardiovascular disease in CKD and non CKD patients [12,13]. Several hypotheses were formulated to explain the association between hypomagnesaemia and impaired AKI recovery. Magnesium relaxes smooth muscle by competing with calcium transport in cell membrane [14]. Cyclosporine and angiotensin II induced contraction in mesangial cells was inhibited by magnesium [15]. In zidovudine treated rats decreased magnesium levels have detrimental effect on GFR and renal blood flow [16]. Magnesium acts as renal vasodilator by acting as Ca²⁺ channel antagonist and also by stimulating release of NO [17]. Magnesium increases RBC deformability, decreases platelet aggregation, has anti-inflammatory action and maintains endothelial cell integrity [18]. Prevalence of hypomagnesaemia varies from 11-65% in different studies [19]. Prevalence was more in ICU patients with sepsis, hypotension, usage of drugs like diuretics and aminoglycosides [2]. Magnesium levels >1.5 meq/L were significantly associated with the non recovery of AKI patients (12 out of 31, p=0.004) on day 1. Studies done previously demonstrated significant association between hypomagnesaemia and AKI non recovery [13,20]. Study done by Alves SC et al., demonstrated that decrease in magnesium levels was significantly associated with the delayed recovery of renal function in AKI patients especially who were seriously ill [20]. In the same study, it was found out that prevalence of hypomagnesaemia was significantly higher in those who had delayed recovery compared with those who recovered from AKI (70% vs. 31%, p=0.003). A study done by Satish R and Gokulnath G showed that decrease in magnesium levels was observed in 62% of patients who were recovering from acute renal failure [21]. About 74% of them had clinical features of hypomagnesaemia and 15 symptomatic patients received intravenous correction. In our study, none of the patients required intravenous magnesium correction. In contrary to the above discussed studies, our observations suggest that patients who were recovered from AKI were more in hypomagnesaemia and normal magnesium levels compared on day 1, 3 and day 6. In patients who were recovered day 1, mean magnesium levels were statistically significant (p=0.001).

The study was done on small sample in a single centre; thus, bias due to non identified centre characteristics was possible and there was also lack of correlation between various causes of AKI and severity of hypomagnesaemia.

Hypomagnesaemia was associated with recovery from AKI (day 1). At the same time magnesium levels >1.5 meq/L on day 1 was also significantly associated with the non recovery of AKI. A greater association with non recovery of AKI was found to be with hypomagnesaemia particularly on day 1 warranting large scale studies to determine the exact role of magnesium in non recovery from AKI.