Magnesiums bronchodilatory effects in the airway have been long accepted. Its effects are attributed to various underlying mechanisms, including an inhibitory effect on acetylcholine release from cholinergic nerve terminals,1 calcium antagonism2 and histamine release from mast cells.3 Magnesium has also been shown to inhibit bronchial smooth muscle contraction by its modulatory effects on calcium channels.3As a result, magnesium has been proposed as a therapeutic option in asthma. Magnesium has been administered via intravenous4-7 and nebulised8-11 routes. The accepted understanding supports the use of intravenous magnesium in acute severe exacerbations of asthma,5,6 but not via the nebulised route.12,13 Although asthma and COPD share some pathophysiological characteristics as well as first-line treatment options in exacerbations,14 there are only a handful of studies investigating magnesium use in COPD.In stable COPD, treatment with intravenous (IV) magnesium has shown to not only reduce lung hyperinflation15,16 but when given with nebulised magnesium, increased forced expiratory volume in one second (FEV1).17Only a handful of studies have investigated the sole use of IV magnesium in acute exacerbations of COPD (AECOPD) and all have supported its use.19-21 One study investigated IV and nebulised magnesium versus nebulised ipratropium and showed similar outcomes in both groups, with slightly greater improvement in spirometry with ipratropium alone.23 One study had investigated nebulised magnesium only and showed no difference in outcome compared to placebo.22 However, the limited amount of evidence, confounding protocols, heterogeneity in doses and timing of administration of magnesium with respect to standard bronchodilators, and conflicting conclusions have led to significant uncertainty regarding the role of magnesium as a treatment option in AECOPD.Our hypothesis was that IV magnesium sulphate administered as an adjunct to standard bronchodilator therapy in AECOPD patients presenting to the Emergency Department (ED) did improve lung function. This pilot study was conducted to assess the above hypothesis, as well as to assess feasibility and effectiveness of the protocol employed for a larger trial.MethodsParticipantsPatients presenting to Palmerston North Hospital ED with a primary diagnosis of AECOPD were invited to participate in this trial between July and October 2013. The period of three months was designated for data collection for this pilot study due to limitations in funding. The diagnosis for AECOPD was made clinically by the attending physician who was not one of the investigators. Clinical symptoms, such as shortness of breath, and signs, such as respiratory rates and wheezing, were used by the clinician to make their diagnosis, but this was not standardised. Patients above the age of 35 years, who had a previously documented diagnosis of COPD by either their general practitioner or in-hospital respiratory specialists were included. Non-infective and infective causes of AECOPD were included. Patients requiring mechanical ventilation or non-invasive ventilation (NIV) at presentation, anyone who was unable to do spirometry or had evidence of pneumothorax or hypotension or any other serious medical condition that would prevent their participation were excluded. Responders or asthma-type COPD patients and those with a history of asthma were excluded by using previously carried out spirometry done at the respiratory outpatient department.Randomisation, masking and ethical approvalPatients were randomly allocated in a double-blind fashion to receive either 2 g IV magnesium sulphate made up to 20 mls in 0.9% sodium chloride solution (saline) or 20 mls of IV saline as placebo. The senior ED pharmacist performed block randomisation and provided identical pre-made syringes with either trial drug or placebo, as per randomisation, to maintain investigator and patient masking. A block size of 20 was used with a 1:1 allocation ratio. Each batch of pre-made syringes expired after seven days and a new batch was made.Ethical approval was granted by the Health and Disability Ethics Committee, New Zealand, approval number 13/NTA/58. The trial was registered withAustralian New Zealand Clinical Trials Registry (ANZCTR) (ACTRN12613000837729).Study ProtocolOn presentation to the ED, all patients were clinically assessed and a venous blood sample was taken to check serum magnesium levels. At this time (TA), Forced Expiratory Volume in 1 second (FEV1) and Forced Vital Capacity (FVC) were measured using a handheld spirometer (EasyOne\u2122, Diagnostic Spirometer. Model number 2001.SN54723/2005. NDD Medical Technologies, Andover, MA 01810, USA). Following this, patients received standard therapy: 5 mg Salbutamol and 500 mcg Ipratropium Bromide by jet nebulisation 60 mg of oral prednisone or 100 mg of IV hydrocortisone Oxygen: 2 litres per minute via nasal prongs if the patients pulse oximetry revealed saturations of <90% Informed consent was obtained by the investigator, who was not the treating clinician, during the first 20 minutes between presentation to the ED and completion of the above standard therapy.Following allocation and immediately after the initial nebulised treatment, a further spirometry was carried out (TB). Immediately after spirometry, all patients received either 2 g of magnesium sulphate or 20 mls of saline only, given via a peripheral vein over 15 minutes. This was given together with another 5 mg of nebulised salbutamol. Spirometry was carried out immediately following the end of the trial drug infusion (T0). A further 5 mg of salbutamol was given via nebulisation at 60 minutes (T60) and 120 minutes (T120) after the trial drug infusion had finished, and spirometry was carried out following each nebulisation. Both groups received the same amount of bronchodilator therapy until T120, after which further treatment was as per the attending physicians clinical judgment.Heart rate, blood pressure, respiratory rate and oxygen saturations by pulse oximetry were monitored at each point to identify adverse reactions. All those performing spirometry had received training from the lung physiology department. Three FEV1 and FVC measurements were made at each time point and the value with two concordant results were used. The spirometer was calibrated weekly using a 3-L syringe provided by the manufacturer, following the manufacturers guidelines. A brief questionnaire was administered at any time after randomisation to determine information regarding medication use, smoking status and flu vaccine uptake. Other interventions, such as further bronchodilator therapies after the spirometry at T120, chest X-rays, antibiotics and analgesia, were at the attending clinicians discretion. The attending clinician also assessed the patient at any time during the trial for the need for NIV or mechanical ventilation and admission into hospital.Statistical analysisAlthough this was a pilot study, a power calculation was performed using an estimated averaged FEV1 of 1.040 litres from a recent trial investigating IV magnesium in AECOPD.21 300 mls of absolute difference in FEV1 between the magnesium and placebo groups was chosen by expert opinion, rather than approximately100 mls suggested by the data from Gonzalez et al (2006). It was unclear what severities of COPD was investigated in the trial.21 Since our trial was open to all severities of COPD, it was conceivable that reversal of bronchoconstriction may be greater compared to the referencing trial.21 Due to a lack of prior data, a standard deviation (SD) of 0.75 litres was chosen by expert opinion. Using these figures for a power of 80% and an alpha error level of 5%, 77 participants were required in each treatment arm, and a sample size of 160 was planned for. Power calculation was carried out using an online calculator.24The primary outcome was the percentage change in FEV1 and FVC at T0, T60 and T120. Baseline lung function is dependent on multiple factors, including age, height and degree of COPD severity. Since our recruitment was open to all severities of COPD, and due to the heterogeneity of lung function expected within each cohorts, mean absolute spirometry values would have been a poor indicator of clinical effect. Investigators believed that the percentage change in spirometry was a better measure to quantify degree of improvement in lung function. However, data was also presented to show the relative differences in the absolute improvements in spirometry at various time points. FEV1 and FVC values from different time points were compared to TB spirometry values, percentage differences were calculated and the means were produced with the SDs.Secondary outcomes were hospital admission, episodes of NIV and/or mechanical ventilation and length of stay. Significance level was set at P< 0.05. A T-test was used to compare the groups at various time points with the baseline at TB (SPSS Statistics, Version 20.0.1. IBM Corporation, 1 New Orchard Road, Armonk, New York, USA). Data was checked for normal distribution and Mann-Whitney U test was applied to non-parametric data.ResultsThirty-seven patients were assessed between July and the end of September, 2013, for eligibility. Three did not meet the inclusion criteria due to other primary causes of shortness of breath and presentation (congestive cardiac failure, asthma). One patient declined to participate in the trial. Thirty-three patients were randomised and 19 were allocated to the placebo group, while 14 were allocated to get magnesium sulphate (Figure 1). Treatment was stopped in one patient in both groups, who were assessed to require NIV therapy shortly after randomisation and prior to the end of the trial drug infusion, hence no further spirometry was carried out on these patients. One patient was excluded from the final analysis in the placebo group due to the patient having previously been enrolled within a period of one week in the magnesium arm of the trial. All other patients received their allocated treatments. All patients included in the final analysis had completed all measurement stages.Figure 1: Trial allocation profileTable 1 details the baseline characteristics of the two groups. There was no difference in mean age, mean pack years, long-term corticosteroid and home nebuliser use. There was no difference in serum magnesium levels, heart rate, respiratory rate, oxygen saturations and mean FEV1 and FVC at presentation (TA). A retrospective analysis revealed that all patients included in the final analysis had a presenting FEV1 of less than 50% predicted.Table 1: Participants baseline characteristics Characteristics Placebo (n=17) Magnesium (n=13) Difference (95% CI) P Value Mean (SD) age, years 72.9 (9.39) 76.1 (12.47) 3.2 (-12.4 to 6.1) 0.48 Female, n (%, 95% CI) 7 (41, 17.6 to 64.4) 2 (15, -4.4 to 34.4) - 0.24 Current smoker, n (%, 95% CI) 5 (30, 8.2 to 51.8) 4 (31, 5.9 to 56.1) - 0.75 Mean (SD) pack years 38.8 (18.2) 40.0 (27.9) 1.2 (-21.9 to 19.3) 0.82 Never smoked, n (%, 95% CI) 1 (6, -5.3 to 17.3) 1 (-6.8 to 22.8) - 0.82 Long term oral steroids, n (%, 95% CI) 1 (6, -5.3 to 17.3) 1 (-6.8 to 22.8) - 0.83 Home nebuliser use, n (%, 95% CI) 1 (6, -5.3 to 17.3) 2 (-4.4 to 34.4) - 0.37 Home oxygen use, n (%, 95% CI) 1 (6, -5.3 to 17.3) 1 (-6.8 to 22.8) - 0.24 Mean (SD) Serum Mg. at presentation, mmol/L 0.78 (0.1) 0.79 (0.1) 0.01 (-0.91 to 0.08) 0.91 Mean (SD) presenting FEV1, mL 691 (288) 637 (293) 54 (-184 to 292) 0.64 Mean (SD) presenting FVC, mL 1770 (719) 1681 (619) 88 (-468 to 644) 0.75 Mean (SD) presenting heart rate 103 (12) 98 (12) 5 (-5.19 to 14.64) 0.34 Mean (SD) Presenting respiratory rate 21 (3) 20 (2) 1 (-2.06 to 2.69) 0.79 Mean (SD) presenting oxygen saturations 91 (4) 90 (4) 1 (-2.71 to 3.44) 0.81 Patients with presenting FEV1<50% predicted, n (%) 17 (100) 13 (100) - - SD: Standard deviation. FEV1: Forced Expiratory Volume in 1 Second. FVC: Forced Vital Capacity. Mg: Magnesium. mL: Millilitres.Mean absolute changes in FEV1 and FVCThere were wide variability of data points indicated by the large standard deviations (Table 2). There was no statistically significant improvement in absolute changes in FEV1 between TA and TB, TB and T0 and T0 and T60. However, there was significantly greater improvement in FEV1 in the magnesium group between 60 minutes and 120 minutes post drug infusion (Table 2). Further, overall, there was significantly greater improvement in FEV1 in the magnesium group at 120 minutes post infusion compared to baseline (166.36 mls vs 80.0mls, Difference of 86.36 mls, P=0.04).Statistically significant improvement in FVC was noted between TB and T0. Similar to FEV1, there was significant improvement in FVC in the magnesium group at 120 minutes post infusion from baseline (333.6 mls vs 149.3 mls, difference of 184.3 mls, P=0.02).Table 2: Mean absolute improvements of FEV1 and FVC at various time intervals FEV1 Time Points Placebo/ mL (SD) Magnesium/ mL (SD) Difference (95% CI) P Value TA-TB 16.0 (108.81) 30.0 (56.57) 14.0 (-60.37 to 88.37) 0.7 TB-T0 18.0 (50.03) 41.8 (53.82) 23.82 (-18.5 to 66.1) 0.26 T0-T60 43.3(66.83) 46.36 (52.21) 3.03 (-47.1 to 53.1) 0.9 T60-T120 18.7 (28.0) 78.18 (84.6) 59.52 (11.47 to 107.56) 0.02 TB-T120 80.0 (102.82) 166.36 (104.72) 86.36 (1.48 to 171.25) 0.04 FVC Placebo/ mL (SD) Magnesium/ mL (SD) Difference (95% CI) P Value TA-TB 20.6 (213.32) 106.36 (162.87) 85.7 (-73.16 to 244.56) 0.28 TB-T0 16.7 (142.61) 140.9 (71.06) 124.2 (27.42 to 221.07) 0.01 T0-T60 36.7 (156.33) 100.9 (84.32) 64.24 (-43.26 to 171.75) 0.23 T60-T120 96.0 (112.49) 91.82 (74.94) 4.18 (-76.60 to 84.96) 0.92 TB-T120 149.3 (223.97) 333.6 (106.14) 184.3 (33.33 to 335.27) 0.02 Time points: TA: at presentation. TB: post initial bronchodilator therapy. T0, T60, T120: Immediately after and at 60 and 120 minutes post-trial drug infusion. SD: Standard deviation. FEV1: Forced Expiratory Volume in 1 Second. FVC: Forced Vital Capacity.Analysis by percentage changeBoth groups presented (TA) with similar FEV1 values (691 mL in placebo vs 637 mL in magnesium group, P= 0.64) (Table 1). The response in FEV1 after standard bronchodilator therapy were similar in both groups (percentage change placebo vs magnesium; 1.52% vs 5.05%, P=0.51). Immediately after trial drug administration (T0), the magnesium group showed greater improvement in FEV1 compared to the placebo group (percentage change of 8.02% vs 2.04%) (Table 3), although this did not attain statistical significance (P=0.06). However, 120 minutes following the administration of the trial drug, the magnesium group showed significantly greater improvement in FEV1 compared to placebo (Table 3).Response in FVC mirrored FEV1 with baseline FVC at TA and percentage change at TB being similar. However at T0 and T60, there were significantly greater improvements in FVC in the magnesium treated patients when compared to TB (Table 3).Table 3: Mean percentage change in FEV1 and FVC Time Points Placebo Magnesium Difference (95% CI) P Value TA, TB 1.52 (15.0) 5.05 (10.3) 3.53 (-7.3 to 14.3) 0.51 TB, T0 2.04 (5.7) 8.02 (10.3) 5.98 (- 0.03 to 12.3) 0.06 TB, T60 8.35 (13.6) 15.87 (14.9) 7.52 (-4.0 to 19.2) 0.19 TB, T120 11.39 (13.6) 27.07 (16.0) 15.68 (3.7 to 27.7) 0.01 Time Points Placebo Magnesium Difference (95% CI) P Value TA, TB -1.24 (14.4) 5.73 (9.6) 6.97 (-3.4 to 17.3) 0.18 TB, T0 2.74 (7.9) 9.80 (7.7) 7.06 (0.7 to 13.5) 0.03 TB, T60 4.59 (13.1) 16.40 (12.1) 11.81 (1.4 to 22.2) 0.03 TB, T120 12.94 (24.5) 22.82 (15.3) 9.88 (-0.5 to 27.3) 0.25 Time points: TA: at presentation. TB: post initial bronchodilator therapy. T0, T60, T120: Immediately after and at 60 and 120 minutes post-trial drug infusion. SD: Standard deviation. FEV1: Forced Expiratory Volume in 1 Second. FVC: Forced Vital Capacity.Secondary outcomesMost participants were admitted into hospital from ED, with one patient in placebo group and two from magnesium group discharged from ED (Table 4). None required mechanical ventilation, whereas one patient was given NIV in the placebo group after all spirometry was completed. Length of hospital stay (LOS) was not significantly lower in the magnesium group (3.18 days compared to 5.47 days, p=0.11). One patient reported hands and facial flushing feeling in the placebo group. No other adverse events were noted by clinicians or reported by patients.Table 4: Secondary outcomes in magnesium and placebo groups Outcome Placebo (n=17) Magnesium (n=13) Difference (95% CI) P Value Requirement for NIV, n (%, 95% CI) 1 (6, -5.3 to 17.3) 0 - 0.3 Requirement for mechanical ventilation, n (%) None None - - Admitted to hospital from ED, n (%, 95% CI) 16 (94, 82.7 to 100) 11 (85, 65.6 to 99.7) - 0.8 Mean (SD) length of hospital stay, days 5.47 (5.03) 3.18 (3.19) 2.28 (-1.28 to 5.85) 0.11 NIV: Non-invasive ventilation. SD: Standard deviation.DiscussionWe had conducted a pilot study for a future, randomised, double-blind, placebo-controlled trial to investigate the effects of IV magnesium as an adjunct therapy to current standard treatments for AECOPD in ED. Our pilot study was the first to investigate the effects of IV magnesium on lung function in AECOPD at a duration longer (120 minutes) than in previous studies (45 minutes).19-21 Thirty patients were included in the final analysis and baseline characteristics were similar in both groups.Data was presented and analysed in two formats to better elucidate an accurate picture; relative differences in absolute improvements in spirometry at each time point, and relative differences in percentage change from baseline. Comparison of absolute change in spirometry is usually the accepted approach. However, given the small participant numbers and the heterogeneity of lung functions within the groups, it would be difficult to detect effects using absolute values for spirometry. Percentage change better quantified lung function changes since it standardised values, which made the results more comparable between cohorts and therefore a better quantification of the effects.Results revealed that there were some significant improvements in FEV1 and FVC in the magnesium group at various time points. Results indicated that IV magnesium seemed to have some immediate beneficial effect on lung function and it may also have a prolonged effect on FEV1 up to two hours post magnesium infusion.This is supported by the handful of studies that have investigated magnesium in AECOPD. Treatment with magnesium seemed to have improved patients symptoms,19 increased PEFR at 45 minutes and almost halved admission rates.20 When given concurrently with salbutamol, intravenous magnesium significantly increased FEV1 compared with control patients.21 One study investigated the combined use of IV and nebulised magnesium against nebulised ipratropium and showed that there was a trend of reduced dyspnoea scores in both groups.23 Furthermore, a retrospective analysis of stable COPD and AECOPD patients revealed that those in the latter group had significantly lower serum magnesium levels.25 Indeed, lower serum magnesium levels seemed to be an independent predictive factor for higher admission rates with acute exacerbations.18Results indicated that the magnesium group had a reduced LOS (3.18 vs 5.47 days), which mirrored another study comparing magnesium with placebo (4.27 vs 7.33 days, p<0.05).20 Our results indicated that a patient presenting with an AECOPD with approximate FEV1 of 700 mL and FVC of 1600 mL, at 2 hours post magnesium infusion, could experience an improvement of 100 mL of FEV1 and 250 mL of FVC greater than just standard therapy. Although this is a clinically significant degree of improvement in lung function in the short term, linking this result to reduced LOS, without supporting spirometry data, would be presumptuous.Indeed this magnitude of bronchodilation is surprising in COPD. We had an all comers approach to recruitment. This was done for two reasons. Firstly, in the acute presentation setting, clinicians do not always know whether it is an infective cause for AECOPD. Secondly, it was unclear from previous studies whether magnesium would have a beneficial effect in all AECOPD severities and hence, no restriction was made regarding severities. Our data revealed a wide variability in baseline spirometry indicating varying severities of AECOPD, although all had a presenting FEV1<50% predicted, indicating all had severe exacerbations. However, no data was collected regarding the severities of these patients baseline COPD. It was conceivable that the magnesium group had patients whose baseline COPD were not as severe as those in the placebo group and hence could have skewed the data.Furthermore, although attempts were made to recruit patients 24 hours a day, there were times when recruitment was not carried out due to lack of investigators or departmental workloads. It was estimated retrospectively that 52 patients had presented with a possible diagnosis of AECOPD over the three months, out of whom 37 were assessed for suitability. It was uncertain whether this could have affected the generalisability of the results, given that no baseline COPD severity data was collected. Future trials will require better definition of patient groups regarding causes of AECOPD and severities of baseline COPD as well as better round-the-clock recruitment.And finally, the original study was planned to be much larger with an n=160, hence a block size of 20 was selected. The authors expected large numbers of patients for the pilot study, using previous years COPD presenters to ED as a guide. Hence the block size of 20 was maintained. Unfortunately, the pilot study was only able to recruit much smaller numbers. The large block size of 20 led to unequal numbers in both cohorts. This may have led to a degree of bias given the small numbers.In conclusion, this pilot trial added to the limited amount of data that indicated that IV magnesium given in conjunction with standard bronchodilator therapy may improve lung function immediately after drug infusion and the effects may have been prolonged for up to two hours post drug infusion. However, future trials will need to ensure adequate patient numbers and stricter patient group definitions. Further, increased duration of spirometry measurements, for example up to 24 hours post magnesium infusion, could be useful in understanding magnesiums effects over prolonged periods.

Magnesiums bronchodilatory effects in the airway have been long accepted. Its effects are attributed to various underlying mechanisms, including an inhibitory effect on acetylcholine release from cholinergic nerve terminals,1 calcium antagonism2 and histamine release from mast cells.3 Magnesium has also been shown to inhibit bronchial smooth muscle contraction by its modulatory effects on calcium channels.3As a result, magnesium has been proposed as a therapeutic option in asthma. Magnesium has been administered via intravenous4-7 and nebulised8-11 routes. The accepted understanding supports the use of intravenous magnesium in acute severe exacerbations of asthma,5,6 but not via the nebulised route.12,13 Although asthma and COPD share some pathophysiological characteristics as well as first-line treatment options in exacerbations,14 there are only a handful of studies investigating magnesium use in COPD.In stable COPD, treatment with intravenous (IV) magnesium has shown to not only reduce lung hyperinflation15,16 but when given with nebulised magnesium, increased forced expiratory volume in one second (FEV1).17Only a handful of studies have investigated the sole use of IV magnesium in acute exacerbations of COPD (AECOPD) and all have supported its use.19-21 One study investigated IV and nebulised magnesium versus nebulised ipratropium and showed similar outcomes in both groups, with slightly greater improvement in spirometry with ipratropium alone.23 One study had investigated nebulised magnesium only and showed no difference in outcome compared to placebo.22 However, the limited amount of evidence, confounding protocols, heterogeneity in doses and timing of administration of magnesium with respect to standard bronchodilators, and conflicting conclusions have led to significant uncertainty regarding the role of magnesium as a treatment option in AECOPD.Our hypothesis was that IV magnesium sulphate administered as an adjunct to standard bronchodilator therapy in AECOPD patients presenting to the Emergency Department (ED) did improve lung function. This pilot study was conducted to assess the above hypothesis, as well as to assess feasibility and effectiveness of the protocol employed for a larger trial.MethodsParticipantsPatients presenting to Palmerston North Hospital ED with a primary diagnosis of AECOPD were invited to participate in this trial between July and October 2013. The period of three months was designated for data collection for this pilot study due to limitations in funding. The diagnosis for AECOPD was made clinically by the attending physician who was not one of the investigators. Clinical symptoms, such as shortness of breath, and signs, such as respiratory rates and wheezing, were used by the clinician to make their diagnosis, but this was not standardised. Patients above the age of 35 years, who had a previously documented diagnosis of COPD by either their general practitioner or in-hospital respiratory specialists were included. Non-infective and infective causes of AECOPD were included. Patients requiring mechanical ventilation or non-invasive ventilation (NIV) at presentation, anyone who was unable to do spirometry or had evidence of pneumothorax or hypotension or any other serious medical condition that would prevent their participation were excluded. Responders or asthma-type COPD patients and those with a history of asthma were excluded by using previously carried out spirometry done at the respiratory outpatient department.Randomisation, masking and ethical approvalPatients were randomly allocated in a double-blind fashion to receive either 2 g IV magnesium sulphate made up to 20 mls in 0.9% sodium chloride solution (saline) or 20 mls of IV saline as placebo. The senior ED pharmacist performed block randomisation and provided identical pre-made syringes with either trial drug or placebo, as per randomisation, to maintain investigator and patient masking. A block size of 20 was used with a 1:1 allocation ratio. Each batch of pre-made syringes expired after seven days and a new batch was made.Ethical approval was granted by the Health and Disability Ethics Committee, New Zealand, approval number 13/NTA/58. The trial was registered withAustralian New Zealand Clinical Trials Registry (ANZCTR) (ACTRN12613000837729).Study ProtocolOn presentation to the ED, all patients were clinically assessed and a venous blood sample was taken to check serum magnesium levels. At this time (TA), Forced Expiratory Volume in 1 second (FEV1) and Forced Vital Capacity (FVC) were measured using a handheld spirometer (EasyOne\u2122, Diagnostic Spirometer. Model number 2001.SN54723/2005. NDD Medical Technologies, Andover, MA 01810, USA). Following this, patients received standard therapy: 5 mg Salbutamol and 500 mcg Ipratropium Bromide by jet nebulisation 60 mg of oral prednisone or 100 mg of IV hydrocortisone Oxygen: 2 litres per minute via nasal prongs if the patients pulse oximetry revealed saturations of <90% Informed consent was obtained by the investigator, who was not the treating clinician, during the first 20 minutes between presentation to the ED and completion of the above standard therapy.Following allocation and immediately after the initial nebulised treatment, a further spirometry was carried out (TB). Immediately after spirometry, all patients received either 2 g of magnesium sulphate or 20 mls of saline only, given via a peripheral vein over 15 minutes. This was given together with another 5 mg of nebulised salbutamol. Spirometry was carried out immediately following the end of the trial drug infusion (T0). A further 5 mg of salbutamol was given via nebulisation at 60 minutes (T60) and 120 minutes (T120) after the trial drug infusion had finished, and spirometry was carried out following each nebulisation. Both groups received the same amount of bronchodilator therapy until T120, after which further treatment was as per the attending physicians clinical judgment.Heart rate, blood pressure, respiratory rate and oxygen saturations by pulse oximetry were monitored at each point to identify adverse reactions. All those performing spirometry had received training from the lung physiology department. Three FEV1 and FVC measurements were made at each time point and the value with two concordant results were used. The spirometer was calibrated weekly using a 3-L syringe provided by the manufacturer, following the manufacturers guidelines. A brief questionnaire was administered at any time after randomisation to determine information regarding medication use, smoking status and flu vaccine uptake. Other interventions, such as further bronchodilator therapies after the spirometry at T120, chest X-rays, antibiotics and analgesia, were at the attending clinicians discretion. The attending clinician also assessed the patient at any time during the trial for the need for NIV or mechanical ventilation and admission into hospital.Statistical analysisAlthough this was a pilot study, a power calculation was performed using an estimated averaged FEV1 of 1.040 litres from a recent trial investigating IV magnesium in AECOPD.21 300 mls of absolute difference in FEV1 between the magnesium and placebo groups was chosen by expert opinion, rather than approximately100 mls suggested by the data from Gonzalez et al (2006). It was unclear what severities of COPD was investigated in the trial.21 Since our trial was open to all severities of COPD, it was conceivable that reversal of bronchoconstriction may be greater compared to the referencing trial.21 Due to a lack of prior data, a standard deviation (SD) of 0.75 litres was chosen by expert opinion. Using these figures for a power of 80% and an alpha error level of 5%, 77 participants were required in each treatment arm, and a sample size of 160 was planned for. Power calculation was carried out using an online calculator.24The primary outcome was the percentage change in FEV1 and FVC at T0, T60 and T120. Baseline lung function is dependent on multiple factors, including age, height and degree of COPD severity. Since our recruitment was open to all severities of COPD, and due to the heterogeneity of lung function expected within each cohorts, mean absolute spirometry values would have been a poor indicator of clinical effect. Investigators believed that the percentage change in spirometry was a better measure to quantify degree of improvement in lung function. However, data was also presented to show the relative differences in the absolute improvements in spirometry at various time points. FEV1 and FVC values from different time points were compared to TB spirometry values, percentage differences were calculated and the means were produced with the SDs.Secondary outcomes were hospital admission, episodes of NIV and/or mechanical ventilation and length of stay. Significance level was set at P< 0.05. A T-test was used to compare the groups at various time points with the baseline at TB (SPSS Statistics, Version 20.0.1. IBM Corporation, 1 New Orchard Road, Armonk, New York, USA). Data was checked for normal distribution and Mann-Whitney U test was applied to non-parametric data.ResultsThirty-seven patients were assessed between July and the end of September, 2013, for eligibility. Three did not meet the inclusion criteria due to other primary causes of shortness of breath and presentation (congestive cardiac failure, asthma). One patient declined to participate in the trial. Thirty-three patients were randomised and 19 were allocated to the placebo group, while 14 were allocated to get magnesium sulphate (Figure 1). Treatment was stopped in one patient in both groups, who were assessed to require NIV therapy shortly after randomisation and prior to the end of the trial drug infusion, hence no further spirometry was carried out on these patients. One patient was excluded from the final analysis in the placebo group due to the patient having previously been enrolled within a period of one week in the magnesium arm of the trial. All other patients received their allocated treatments. All patients included in the final analysis had completed all measurement stages.Figure 1: Trial allocation profileTable 1 details the baseline characteristics of the two groups. There was no difference in mean age, mean pack years, long-term corticosteroid and home nebuliser use. There was no difference in serum magnesium levels, heart rate, respiratory rate, oxygen saturations and mean FEV1 and FVC at presentation (TA). A retrospective analysis revealed that all patients included in the final analysis had a presenting FEV1 of less than 50% predicted.Table 1: Participants baseline characteristics Characteristics Placebo (n=17) Magnesium (n=13) Difference (95% CI) P Value Mean (SD) age, years 72.9 (9.39) 76.1 (12.47) 3.2 (-12.4 to 6.1) 0.48 Female, n (%, 95% CI) 7 (41, 17.6 to 64.4) 2 (15, -4.4 to 34.4) - 0.24 Current smoker, n (%, 95% CI) 5 (30, 8.2 to 51.8) 4 (31, 5.9 to 56.1) - 0.75 Mean (SD) pack years 38.8 (18.2) 40.0 (27.9) 1.2 (-21.9 to 19.3) 0.82 Never smoked, n (%, 95% CI) 1 (6, -5.3 to 17.3) 1 (-6.8 to 22.8) - 0.82 Long term oral steroids, n (%, 95% CI) 1 (6, -5.3 to 17.3) 1 (-6.8 to 22.8) - 0.83 Home nebuliser use, n (%, 95% CI) 1 (6, -5.3 to 17.3) 2 (-4.4 to 34.4) - 0.37 Home oxygen use, n (%, 95% CI) 1 (6, -5.3 to 17.3) 1 (-6.8 to 22.8) - 0.24 Mean (SD) Serum Mg. at presentation, mmol/L 0.78 (0.1) 0.79 (0.1) 0.01 (-0.91 to 0.08) 0.91 Mean (SD) presenting FEV1, mL 691 (288) 637 (293) 54 (-184 to 292) 0.64 Mean (SD) presenting FVC, mL 1770 (719) 1681 (619) 88 (-468 to 644) 0.75 Mean (SD) presenting heart rate 103 (12) 98 (12) 5 (-5.19 to 14.64) 0.34 Mean (SD) Presenting respiratory rate 21 (3) 20 (2) 1 (-2.06 to 2.69) 0.79 Mean (SD) presenting oxygen saturations 91 (4) 90 (4) 1 (-2.71 to 3.44) 0.81 Patients with presenting FEV1<50% predicted, n (%) 17 (100) 13 (100) - - SD: Standard deviation. FEV1: Forced Expiratory Volume in 1 Second. FVC: Forced Vital Capacity. Mg: Magnesium. mL: Millilitres.Mean absolute changes in FEV1 and FVCThere were wide variability of data points indicated by the large standard deviations (Table 2). There was no statistically significant improvement in absolute changes in FEV1 between TA and TB, TB and T0 and T0 and T60. However, there was significantly greater improvement in FEV1 in the magnesium group between 60 minutes and 120 minutes post drug infusion (Table 2). Further, overall, there was significantly greater improvement in FEV1 in the magnesium group at 120 minutes post infusion compared to baseline (166.36 mls vs 80.0mls, Difference of 86.36 mls, P=0.04).Statistically significant improvement in FVC was noted between TB and T0. Similar to FEV1, there was significant improvement in FVC in the magnesium group at 120 minutes post infusion from baseline (333.6 mls vs 149.3 mls, difference of 184.3 mls, P=0.02).Table 2: Mean absolute improvements of FEV1 and FVC at various time intervals FEV1 Time Points Placebo/ mL (SD) Magnesium/ mL (SD) Difference (95% CI) P Value TA-TB 16.0 (108.81) 30.0 (56.57) 14.0 (-60.37 to 88.37) 0.7 TB-T0 18.0 (50.03) 41.8 (53.82) 23.82 (-18.5 to 66.1) 0.26 T0-T60 43.3(66.83) 46.36 (52.21) 3.03 (-47.1 to 53.1) 0.9 T60-T120 18.7 (28.0) 78.18 (84.6) 59.52 (11.47 to 107.56) 0.02 TB-T120 80.0 (102.82) 166.36 (104.72) 86.36 (1.48 to 171.25) 0.04 FVC Placebo/ mL (SD) Magnesium/ mL (SD) Difference (95% CI) P Value TA-TB 20.6 (213.32) 106.36 (162.87) 85.7 (-73.16 to 244.56) 0.28 TB-T0 16.7 (142.61) 140.9 (71.06) 124.2 (27.42 to 221.07) 0.01 T0-T60 36.7 (156.33) 100.9 (84.32) 64.24 (-43.26 to 171.75) 0.23 T60-T120 96.0 (112.49) 91.82 (74.94) 4.18 (-76.60 to 84.96) 0.92 TB-T120 149.3 (223.97) 333.6 (106.14) 184.3 (33.33 to 335.27) 0.02 Time points: TA: at presentation. TB: post initial bronchodilator therapy. T0, T60, T120: Immediately after and at 60 and 120 minutes post-trial drug infusion. SD: Standard deviation. FEV1: Forced Expiratory Volume in 1 Second. FVC: Forced Vital Capacity.Analysis by percentage changeBoth groups presented (TA) with similar FEV1 values (691 mL in placebo vs 637 mL in magnesium group, P= 0.64) (Table 1). The response in FEV1 after standard bronchodilator therapy were similar in both groups (percentage change placebo vs magnesium; 1.52% vs 5.05%, P=0.51). Immediately after trial drug administration (T0), the magnesium group showed greater improvement in FEV1 compared to the placebo group (percentage change of 8.02% vs 2.04%) (Table 3), although this did not attain statistical significance (P=0.06). However, 120 minutes following the administration of the trial drug, the magnesium group showed significantly greater improvement in FEV1 compared to placebo (Table 3).Response in FVC mirrored FEV1 with baseline FVC at TA and percentage change at TB being similar. However at T0 and T60, there were significantly greater improvements in FVC in the magnesium treated patients when compared to TB (Table 3).Table 3: Mean percentage change in FEV1 and FVC Time Points Placebo Magnesium Difference (95% CI) P Value TA, TB 1.52 (15.0) 5.05 (10.3) 3.53 (-7.3 to 14.3) 0.51 TB, T0 2.04 (5.7) 8.02 (10.3) 5.98 (- 0.03 to 12.3) 0.06 TB, T60 8.35 (13.6) 15.87 (14.9) 7.52 (-4.0 to 19.2) 0.19 TB, T120 11.39 (13.6) 27.07 (16.0) 15.68 (3.7 to 27.7) 0.01 Time Points Placebo Magnesium Difference (95% CI) P Value TA, TB -1.24 (14.4) 5.73 (9.6) 6.97 (-3.4 to 17.3) 0.18 TB, T0 2.74 (7.9) 9.80 (7.7) 7.06 (0.7 to 13.5) 0.03 TB, T60 4.59 (13.1) 16.40 (12.1) 11.81 (1.4 to 22.2) 0.03 TB, T120 12.94 (24.5) 22.82 (15.3) 9.88 (-0.5 to 27.3) 0.25 Time points: TA: at presentation. TB: post initial bronchodilator therapy. T0, T60, T120: Immediately after and at 60 and 120 minutes post-trial drug infusion. SD: Standard deviation. FEV1: Forced Expiratory Volume in 1 Second. FVC: Forced Vital Capacity.Secondary outcomesMost participants were admitted into hospital from ED, with one patient in placebo group and two from magnesium group discharged from ED (Table 4). None required mechanical ventilation, whereas one patient was given NIV in the placebo group after all spirometry was completed. Length of hospital stay (LOS) was not significantly lower in the magnesium group (3.18 days compared to 5.47 days, p=0.11). One patient reported hands and facial flushing feeling in the placebo group. No other adverse events were noted by clinicians or reported by patients.Table 4: Secondary outcomes in magnesium and placebo groups Outcome Placebo (n=17) Magnesium (n=13) Difference (95% CI) P Value Requirement for NIV, n (%, 95% CI) 1 (6, -5.3 to 17.3) 0 - 0.3 Requirement for mechanical ventilation, n (%) None None - - Admitted to hospital from ED, n (%, 95% CI) 16 (94, 82.7 to 100) 11 (85, 65.6 to 99.7) - 0.8 Mean (SD) length of hospital stay, days 5.47 (5.03) 3.18 (3.19) 2.28 (-1.28 to 5.85) 0.11 NIV: Non-invasive ventilation. SD: Standard deviation.DiscussionWe had conducted a pilot study for a future, randomised, double-blind, placebo-controlled trial to investigate the effects of IV magnesium as an adjunct therapy to current standard treatments for AECOPD in ED. Our pilot study was the first to investigate the effects of IV magnesium on lung function in AECOPD at a duration longer (120 minutes) than in previous studies (45 minutes).19-21 Thirty patients were included in the final analysis and baseline characteristics were similar in both groups.Data was presented and analysed in two formats to better elucidate an accurate picture; relative differences in absolute improvements in spirometry at each time point, and relative differences in percentage change from baseline. Comparison of absolute change in spirometry is usually the accepted approach. However, given the small participant numbers and the heterogeneity of lung functions within the groups, it would be difficult to detect effects using absolute values for spirometry. Percentage change better quantified lung function changes since it standardised values, which made the results more comparable between cohorts and therefore a better quantification of the effects.Results revealed that there were some significant improvements in FEV1 and FVC in the magnesium group at various time points. Results indicated that IV magnesium seemed to have some immediate beneficial effect on lung function and it may also have a prolonged effect on FEV1 up to two hours post magnesium infusion.This is supported by the handful of studies that have investigated magnesium in AECOPD. Treatment with magnesium seemed to have improved patients symptoms,19 increased PEFR at 45 minutes and almost halved admission rates.20 When given concurrently with salbutamol, intravenous magnesium significantly increased FEV1 compared with control patients.21 One study investigated the combined use of IV and nebulised magnesium against nebulised ipratropium and showed that there was a trend of reduced dyspnoea scores in both groups.23 Furthermore, a retrospective analysis of stable COPD and AECOPD patients revealed that those in the latter group had significantly lower serum magnesium levels.25 Indeed, lower serum magnesium levels seemed to be an independent predictive factor for higher admission rates with acute exacerbations.18Results indicated that the magnesium group had a reduced LOS (3.18 vs 5.47 days), which mirrored another study comparing magnesium with placebo (4.27 vs 7.33 days, p<0.05).20 Our results indicated that a patient presenting with an AECOPD with approximate FEV1 of 700 mL and FVC of 1600 mL, at 2 hours post magnesium infusion, could experience an improvement of 100 mL of FEV1 and 250 mL of FVC greater than just standard therapy. Although this is a clinically significant degree of improvement in lung function in the short term, linking this result to reduced LOS, without supporting spirometry data, would be presumptuous.Indeed this magnitude of bronchodilation is surprising in COPD. We had an all comers approach to recruitment. This was done for two reasons. Firstly, in the acute presentation setting, clinicians do not always know whether it is an infective cause for AECOPD. Secondly, it was unclear from previous studies whether magnesium would have a beneficial effect in all AECOPD severities and hence, no restriction was made regarding severities. Our data revealed a wide variability in baseline spirometry indicating varying severities of AECOPD, although all had a presenting FEV1<50% predicted, indicating all had severe exacerbations. However, no data was collected regarding the severities of these patients baseline COPD. It was conceivable that the magnesium group had patients whose baseline COPD were not as severe as those in the placebo group and hence could have skewed the data.Furthermore, although attempts were made to recruit patients 24 hours a day, there were times when recruitment was not carried out due to lack of investigators or departmental workloads. It was estimated retrospectively that 52 patients had presented with a possible diagnosis of AECOPD over the three months, out of whom 37 were assessed for suitability. It was uncertain whether this could have affected the generalisability of the results, given that no baseline COPD severity data was collected. Future trials will require better definition of patient groups regarding causes of AECOPD and severities of baseline COPD as well as better round-the-clock recruitment.And finally, the original study was planned to be much larger with an n=160, hence a block size of 20 was selected. The authors expected large numbers of patients for the pilot study, using previous years COPD presenters to ED as a guide. Hence the block size of 20 was maintained. Unfortunately, the pilot study was only able to recruit much smaller numbers. The large block size of 20 led to unequal numbers in both cohorts. This may have led to a degree of bias given the small numbers.In conclusion, this pilot trial added to the limited amount of data that indicated that IV magnesium given in conjunction with standard bronchodilator therapy may improve lung function immediately after drug infusion and the effects may have been prolonged for up to two hours post drug infusion. However, future trials will need to ensure adequate patient numbers and stricter patient group definitions. Further, increased duration of spirometry measurements, for example up to 24 hours post magnesium infusion, could be useful in understanding magnesiums effects over prolonged periods.