Effect of preoperative bicarbonate infusion on maternal and perinatal outcomes of obstructed labour in Mbale Regional Referral Hospital: A study protocol for a randomised controlled trial

Introduction To improve maternal and fetal outcomes among patients with obstructed labour (OL) in low-resource settings, the associated electrolyte and metabolic derangements must be adequately corrected. Oral fluid intake during labour and preoperative intravenous fluid replacement following OL corrects the associated dehydration and electrolyte changes, but it does not completely reverse the metabolic acidosis, that is, a cause of intrapartum birth asphyxia and a risk factor for primary postpartum haemorrhage due to uterine atony. Sodium bicarbonate is a safe, effective, cheap and readily available acid buffer, that is widely used by sportspeople to improve performance. It also appears to improve fetal and maternal outcomes in abnormally progressing labour. However, its effects on maternal and fetal outcomes among patients with OL is unknown. We aim at establishing the effect of a single-dose preoperative infusion of sodium bicarbonate on maternal and fetal lactate levels and clinical outcomes among patients with OL. Methods and analysis This will be a double blind, randomised controlled clinical phase IIb trial. We will randomise 478 patients with OL to receive either 50 mL of placebo with standard preoperative infusion of normal saline (1.5 L) or 4.2 g of sodium bicarbonate solution (50 mL of 50 mmol/L) with the preoperative infusion of normal saline (1.5 L). The primary outcome will be mean lactate levels in maternal capillary blood at 1 hour after study drug administration and in the arterial cord blood at birth. We will use the intention-to-treat analysis approach. Secondary outcomes will include safety, maternal and fetal morbidity and mortality up to 14 days postpartum. Ethics and dissemination Makerere University School of Medicine Research and Ethics Committee and Uganda National Council for Science and Technology have approved the protocol. Each participant will give informed consent at enrollment. Trial registration number PACTR201805003364421 This will be a superiority, double blind, randomised controlled clinical phase IIb trial. Half of the 478 patients with OL will receive the intervention (sodium bicarbonate infusion) with preoperative normal saline infusion, and the other half will receive the standard of care (preoperative saline infusion) alone. We will conduct this study in Mbale Regional Referral Hospital located at the heart of Mbale Municipality, 214 km to the East of the capital city, Kampala. It is the main referral hospital, serving 14 districts in the Elgon zone, bordering western Kenya. This is a government run, not-for-profit, charge-free, 470-bed hospital with 52 maternity beds. The Department of Obstetrics and Gynaecology has one consultant, two specialists, three medical officers and 21 Midwives. In terms of outputs, the labour and delivery suite is only second to the Mulago National Referral Hospital labour suite. Annually, about 12 000 childbirths occur in this hospital with a caesarean section rate of 35% and ~500 mothers have OL. The MOH has ranked it as the best performing Regional Referral Hospital in Uganda for the last 4 years. This study will be carried out among patients with OL admitted to the labour suite in Mbale Regional Referral Hospital for emergency caesarean section during the period of the study. Either a medical officer or specialist on duty using the American Association of Obstetricians and Gynaecologists (ACOG) definition will make the diagnosis of OL. In the first stage of labour, she should have cervical dilatation >6 cm with ruptured membranes, adequate contractions lasting >4 hours with no change in cervical dilatation or delay in the second active stage of labour (nullipara >2 hours, multipara >1 hour) with adequate uterine contractions. In addition, any two of: the obvious signs of severe obstruction such as caput formation, severe moulding, Bandl’s ring, subconjunctival haemorrhages or an oedematous vulva. We will include patients with OL carrying singleton, term pregnancies (≥37 weeks of gestation) in cephalic presentation. We will exclude patients with other obstetric emergencies such as antepartum haemorrhage, preeclampsia and eclampsia (defined as elevated blood pressure of at least 140/90 mm Hg, urine protein of at least 2+, any of the danger signs and fits), premature rupture of membranes and intrauterine fetal death. Patients with comorbidities such as diabetes mellitus, sickle cell disease, renal disease, liver disease and heart disease. We will also exclude those patients with hypokalemia (<3.3 mmol/L), hypocalcaemia (<8.2 mmol/L), hypernatraemia (>148 mmol/L) and alkalosis (bicarbonate >22 mmol/L) because they are more likely to develop adverse drug reactions. An independent biostatician will generate a sequence of random numbers using the online randomisation service of www.sealedenvelope.com in permuted block sizes of four, six and eight. Based on this sequence, OL patients will be randomly allocated to either intervention or control arms in a 1:1 ratio. An independent pharmacist who is not involved in the recruitment of study participants, will conceal the randomisation sequence by preparing new labels with sequential numbers to be placed on identical study drug packages each containing five similar 10 mL glass vials without the original labels. After consent for inclusion is confirmed, a study nurse will take the next study drug package and administer its contents to the participant. The intervention will be a preoperative infusion of 50 mL of sodium bicarbonate 8.4% solution equivalent to 4.2 g or 50 mmol/L of bicarbonate (Martindale Pharma, Essex, UK) in 10 mL glass vials. The sodium bicarbonate will be administered intravenously as a bolus immediately after recruitment by trained research assistants who are all experienced midwives working in the labour suite, followed by 1.5 L of normal saline over the next hour. Participants in the control arm will receive a preoperative infusion of normal saline, which is part of the current standard of care. Fifty millilitres of sodium chloride 0.9% in identical 10 mL glass vials (AccuHealth Care, Gujarat, India) will be administered intravenously as a bolus immediately after recruitment by trained research assistants, who are all experienced midwives working in the labour suite, followed by 1.5 L normal saline over the next hour. In addition, recruits will receive the standard preoperative care which includes preoperative antibiotic prophylaxis, at least 1.5 L of intravenous fluids preoperatively, bladder emptying, administration of oxygen and lying in left-lateral position.16 The primary outcomes in this study will be the mean lactate levels in maternal capillary blood at 1 hour after the onset of study drug administration and in arterial cord blood within 1 min of birth. We will measure lactate at the bedside using a hand-held Lactate Pro2 device (Arkray Factory). The secondary maternal outcomes will include myometrial lactate levels at caesarean section, maternal mortality up to 14 days postpartum. Other morbidities such as primary PPH, birth canal injuries, duration of admission, puerperal sepsis (persistent fever >38°C, chills and general malaise, pain in the lower abdomen, persistent bloody/pus discharge [lochia] from genital tract, which may have an unpleasant smell, tenderness on palpating the uterus uterine subinvolution, wound dehiscence/burst abdomen),16 fistulae and readmissions. Secondary perinatal outcomes will include mean lactate in venous cord blood, APGAR score, admission to the neonatal intensive care unit (NICU), asphyxia, neonatal sepsis (irritability, poorly breast feeding, bulging anterior fontanelle)16 and perinatal death up to 7 days postpartum. For the secondary safety outcomes, we will monitor for the following drug reactions throughout the period of the study. Frequent urge to urinate, continuing headache, continuing loss of appetite, mood or mental changes, muscle pain or twitching, nausea or vomiting, stomach crumps, slow breathing, swelling of feet or lower limbs, unpleasant taste, increased thirst, unusual tiredness or weakness, venous irritation, cellulitis and intravenous site pain.20–22 Using an interviewer administered questionnaire and available records (antenatal cards, facility registers and case report files), sociodemographic and clinical characteristics will be collected by trained research assistants. At baseline, five millilitres of blood will be collected in the appropriate vacutainers for a complete blood count, renal function tests, liver function tests and electrolytes. Ten millilitres of fresh urine will be collected in a sterile container for analysis. All these specimens will be delivered to MBN clinical laboratories within 1 hour of collection for analysis. Figure 1 shows the Consolidated Standards of Reporting Trials flow diagram for this study. We will follow-up patients for up to 14 days postpartum either by phone call if they are discharged or by direct visits if they are still admitted according to the current standard of care for patients with OL in Uganda. Flow diagram of study participants in the trial. Since the multiple testing in this study will be corrected for at analysis, we estimated the sample size to cater for that. The testing will take place c = two times, the final analysis will use a Bonferroni-Holm correction to adjust for multiplicity and the critical value will divide the α level by c. To detect a Δ=15% difference in mean lactate levels between the intervention and control arms. Assuming an equal number of participants in each group, a two-sided significance level α of 0.05 for a 95% CI, a power (1−β) of 90%, an allowance of c = two multiple comparisons and a Bonferroni-Holm method for comparison of means, we used the formula, n>=(Z1−α2c+Z1−β)22(SDΔ)2 23 and Open Epi24 to determine the sample size. Where Z1−α2c=Z1−0.052∗2=Z1−0.0125=2.24 and Z1−β=Z1−0.1=Z0.9=1.28 using standard normal tables. The mean and SD of the maternal venous lactate at the end of second stage of labour is 2.6±1.0 mmol/L25 without any use of bicarbonate. To detect a difference of 15% (0.39 mmol/L) at 1 hour and assuming the same SD of ±1 in both arms, 326 participants will be required. Correcting for an attrition rate of 10%,14 gives a total sample size of 364. The mean arterial umbilical cord blood lactate at 37 weeks of gestation is 4.3±1.9 mmol/L26 without any use of bicarbonate. In order to detect a difference of 15% (0.645 mmol/L) at birth and assuming the same SD of ±1.9 in both groups, 432 participants will be recruited. Correcting for an attrition rate of 10%14 gives a total sample size of 478. We therefore chose a sample size of 478 to provide adequate power for both hypotheses. Well-trained research assistants (RAs) will collect data using a pretested interviewer administered electronic questionnaire on password protected smart phones using the Open Data Kit software.27 To increase accuracy, the data will be triangulated with a review of relevant health facility records such as the antenatal cards, the maternity and theatre registers and the participants’ case notes. The questionnaire will be coded with checks for internal consistency. Data on sociodemographic, clinical and laboratory parameters will be collected at baseline, at 1 hour after onset of study drug administration for the primary maternal outcome, at the time of childbirth for the primary neonatal outcome and at 7 and 14 days postpartum for the secondary perinatal and maternal outcomes as summarised in table 1. The principal investigator (PI) will review the entries from the Google aggregate server every 24 hours to ensure data quality and completeness. Summary of the study procedures and timelines This will be conducted using STATA V.14 software or higher using the principle of ‘intention to treat’. Descriptive statistics will be used to summarise baseline characteristics of the study participants and assess if randomisation was successful. The primary maternal outcome will be the difference in capillary blood lactate levels at 1 hour after onset of study drug administration. The Bonferroni-Holm method28 29 will be used to compare the difference in means at baseline and 1 hour after onset of study drug administration both within and between each of the two arms. For the primary fetal outcome of mean arterial cord blood lactate at birth, the Bonferroni-Holm method will be used to compare the mean lactate levels in the two arms within 1 min of childbirth. The secondary outcomes will be reported as proportions in each of the two arms. Categorical variables will be compared using the χ2 and Fisher’s exact tests. Potential confounders and effect modifiers unbalanced at baseline and associated with the outcome (p<0.05) will be adjusted for using multivariable linear or/and logistic regression. Proportions and the number needed to treat/number needed to harm will be reported for the secondary maternal and fetal outcomes. We will do a subgroup analysis of patients admitted as referrals with a diagnosis of OL (to represent patients most likely to have neglected OL) to compare them with those who are diagnosed with OL within the referral hospital. A second subgroup analysis will be for those patients who give birth >2 hours after administration of the study drug, when we expect the effect of the intervention to have worn off. We will conduct a dry run for a period of 1 month before introducing the intervention. To facilitate the training of all the research assistants in the study protocol procedures, filling of study questionnaires using the open data kit (ODK) software,27 accurate measurement of lactate at the bedside using the Lactate Pro2 device and the ideal technique for collection of samples especially blood to avoid haemolysis. The MBN clinical laboratories are internationally accredited and they are involved in regular internal and external quality control checks. The sponsor of this study (Makerere University) does not formally monitor studies to ensure compliance and adherence to the standard operating procedures of the study protocol. The PI will check each case report form on submission for completeness and undertake regular interviews with study staff and a sample of the study participants to check on the adherence. In addition, the regulatory bodies such as the Institutional review board (IRB), Uganda National Council for Science and Technology (UNCST) and National Drug Authority also carry out regular scheduled and unscheduled spot checks to monitor adherence of the study to the approved protocol. During the study, all the serious adverse events will be actively identified and reported to the IRB within 24 hours of occurrence. We will adopt and use the School of Medicine Research Ethics Committee reporting form. Only qualified health workers will be recruited and trained in the protocol to work as research assistants on this trial. The independent data monitoring committee will review unblinded data when one-third of the participants have been enrolled and followed up to completion and report to the sponsor of the study. If need arises such as patient safety, the study Steering Committee and the Independent data monitoring committee (IDMC) will request the independent study biostatistician to unblind the treatment allocation for a specific patient or group of patients without compromising the allocation concealment for the rest of the participants. Results will be disseminated to the study participants through the local radio stations and local council community meetings at the village level. Findings will be shared with colleagues and administrators in Mbale Regional Referral Hospital/Busitema University Faculty of Health Sciences and the Uganda MOH through workshops and seminars. To reach the wider scientific community, the findings will be published in open access peer-reviewed journals and presented at both local and international conferences. The datasets will be provided free of charge by the primary author on request. The patients and public were not involved in the design and conceptualisation of this study.