Ethical clearance to conduct this study was obtained from the University of the Witwatersrand and Human Research Ethics Committee (No. MED24-08-553).

Preoperatively, all patients met the criteria for bariatric surgery. According to the South African Society for Surgery, Obesity and Metabolism guidelines, patients may be referred for surgical management only after assessment by a specialist weight management service.¹¹ Patients who meet the criteria for bariatric surgery are individuals who fit the following criteria: individuals without any clinical or psychological contraindications; patients with a BMI above 40 kg/m²; patients with a BMI above 35 kg/m² with comorbid disease where conventional treatment has failed; and patients with a BMI > 30 kg/m² with poorly controlled diabetes mellitus.¹² In our study, the median BMI was 59, and based on this criterion alone, all patients qualified for surgery.

This figure is significantly higher than the global average for patients presenting for bariatric surgery.⁹ Since 2014, the IFSO has produced an annual report detailing all bariatric surgeries submitted to a global registry. In its 2024 report, which marked the eighth yearly publication, results from 5 million surgeries performed in 19 countries worldwide were tabulated. South Africa had the highest BMI for males at 47.7, based on 362 surgeries performed in 2024 (across both public and private settings).⁹ This contrasts with the global average BMI of 40 kg/m² – 45 kg/m². However, in keeping with international trends, our cohort was over 90% female.⁹

Specialist services included in our cohort were a dietitian and a psychologist to aid weight loss, but these were provided in only 72% of cases. All patients achieved some weight loss; the average was 6.7% of their initial weight. However, only 65.1% achieved the 5% – 10% preoperative weight loss, which ERAS recommends.⁶

Surgical optimisation encompasses education and counselling, smoking cessation, alcohol abstinence and preoperative weight loss. Preoperative education was well provided by the surgical team and the psychologists. This was not formally assessed in terms of patient understanding.

During the audit, it was found that specific preoperative recommendations were suboptimal in accordance with ERAS guidelines.⁶ Improved outcomes and adherence to the ERAS protocol have been observed in other lower- and middle-income countries that adopt a multidisciplinary team approach.^13,14 In a study published in Brazil by Chaim et al.,¹³ the adoption of a multidisciplinary team approach in the preoperative period resulted in greater weight loss, reversal of patient comorbidities and fewer postoperative complications. Currently at HJH, all units involved in bariatric surgery operate in isolation: dietetics, surgery, endocrinology, psychology and anaesthesia. A holistic multidisciplinary approach for patients at HJH may improve perioperative outcomes.

Airway management was ubiquitously performed via endotracheal intubation. This is in keeping with ERAS guidelines, as these patients are at an increased risk of aspiration due to surgical and patient factors.¹⁵ However, ERAS recommendations advocate ramping the patient to improve laryngeal view and oxygenation on induction⁶; this was done in only 30.2% of patients. The low implementation of a ramped-up position in the cohort may reflect limited access to ramping pillow devices, a lack of bariatric airway protocols, and variability in familiarity with bariatric airway optimisation techniques. Neuromuscular blockade was used in 95% of patients; however, train-of-four (TOF) monitoring was performed in only one patient. The guidelines recommend deep neuromuscular blockade intraoperatively, but this was not monitored. This is problematic, as the protocol advocates objective neuromuscular monitoring, in which a TOF > 0.9 is achieved before extubation.⁶ Residual neuromuscular blockade increases the risk of airway obstruction and pulmonary complications.⁶ The low utilisation of TOF monitoring observed in our study likely reflects a combination of equipment limitations, gaps in training, absence of a standardised anaesthetic protocol and reliance on subjective clinical assessment rather than objective neuromuscular monitoring.

In terms of analgesia, ERAS recommendations strongly suggest multimodal analgesia incorporating non-opioid agents, such as non-steroidal anti-inflammatories (NSAIDs), acetaminophen and regional techniques to minimise opioid side effects and improve recovery.⁶ Common adjuncts include lignocaine, dexmedetomidine, ketamine and magnesium sulphate.⁶ Opioids should only be administered as a rescue protocol at the lowest effective dose and should be short-acting agents like fentanyl.⁶ In our study, patients received at least one opioid, the most common opioid being remifentanil (76.7%; n = 33/43), followed by fentanyl (60.5%; n = 26/43), and morphine (44.2%; n = 19/43). Importantly, 55% of the patients had opioid-sparing regimes. These regimens consisted of ERAS-recommended drugs. A multimodal approach was used in most patients. Ketamine, an N-methyl-D-aspartate (NMDA) receptor antagonist, was administered to 34 patients (79.1%). Lignocaine was used in 36 patients (83.7%). An alpha-2 agonist, dexmedetomidine, was used in 32 patients (74.4%). Paracetamol was administered in 31 patients (72.1%), whereas NSAIDs were used less frequently, in 10 patients (23.2%).

Specific analgesic vary by institution. In a systematic review looking at the efficacy of non-opioid analgesics and adjuvants in multimodal analgesia for reducing postoperative opioid consumption and complications in obesity, it was found that the use of dexmedetomidine, ketamine, lignocaine, magnesium and gabapentin significantly reduced the use of postoperative opioids, PONV and length of hospital stay.¹⁶ In a speciality bariatric hospital in Hong Kong, a protocol was instituted enlisting opioid-sparing analgesia strategies.¹⁷ Preoperatively, paracetamol and pregabalin were given. After induction, a dexmedetomidine infusion using lean body weight and fentanyl was administered. Maintenance included a dexmedetomidine infusion, a ketamine bolus, parecoxib, tramadol and local infiltration of the wound. Postoperative analgesia included paracetamol, etoricoxib and pregabalin. Rescue analgesia was pethidine. The study noted, however, that 46% of patients required pethidine in the postoperative period.¹⁷

Opioid minimisation has been shown to reduce PONV, improve recovery times and shorten hospital stays.⁶ Notably, all of the patients in the study received at least one opioid. Morphine was administered to 44.2% of patients. Morphine is not recommended as a first-line treatment in bariatric surgery because of the risk of respiratory and gastrointestinal complications, slower recovery, and a higher potential for long-term opioid dependence.¹⁸ In this study population, 18% of the group were diagnosed with obstructive sleep apnoea (OSA), and the majority had significant risk factors for OSA. Therefore, the use of morphine can be problematic because of the risk of respiratory depression.¹⁹

The significant use of morphine in these patients could be attributed to the lack of analgesia protocols in the unit and the limited availability of drugs. However, morphine has been used safely after bariatric surgery. In a prospective randomised controlled study, the use of morphine in an intrathecal route using 0.3 mg added to bupivacaine 0.5%, with a multimodal analgesic plan, was reported to be effective in reducing postoperative pain scores in patients who underwent bariatric surgery.²⁰ There were no statistically significant respiratory depression or respiratory complications secondary to opioid use found in this study.

The ERAS guidelines for bariatric surgery recommend multimodal prophylaxis against PONV. All patients should receive two or more antiemetics, including a steroid and a serotonin receptor antagonist, and consider a third agent in patients who have a very high risk for PONV. The Apfel score can be inadequate for guiding practitioners in determining which patients are at very high risk.²¹ There was a high adherence to this recommendation in this study, with 78% of patients receiving at least two agents. Monotherapy was used in 25.6% of patients.

In a prospective study that analysed more than 8000 patients, it was found that patients who were at very high risk for PONV were more likely to be younger, female and undergoing laparoscopic sleeve gastrectomy with moderate to severe pain postoperatively.²² Recommended additional agents include a neurokinin-1 receptor antagonist or a transdermal scopolamine patch.⁶ Neither of these agents is available at HJH.

Target-controlled infusion was used in 72.1% of patients; however, the model used was not noted. Only 58.1% were monitored for depth of anaesthesia. There was one case of postoperative awareness, which may be attributable to the above. The use of TCI should be standardised, as the incidence of PONV has been shown to decrease in propofol-based TCI regimes.²³ Target-controlled infusion has been well corroborated in the benefits of PONV in this population; however, in a systematic review and meta-analysis, there was no difference in postoperative pain scores, postoperative opioid use and perioperative complications compared to their volatile counterparts.²⁴ However, as a result of the altered pharmacodynamics in obese individuals, the need for accurate dosing and the potential for infusion interruptions, it should be mandatory that depth of anaesthesia monitoring be used to minimise the risk of awareness.²⁵

There are no specific ERAS protocols for the use of TIVA/TCI in bariatric surgery, nor does it mandate its use over volatile anaesthesia.⁶ However, it indirectly supports the use of TIVA/TCI to prevent PONV. In the demographic of obese individuals, the Eleveld model for propofol should be used as it has the strongest validation across BMI ranges.²⁶ This model has been shown to accurately estimate the clearance in obese patients, which is essential for precise dosing in this population.²⁷ When the Eleveld model is unavailable, the Marsh or Schneider method can be used. The input weight should be the adjusted body weight for the Marsh model or fat-free mass for the Schneider model.²⁸ These models should be used in conjunction with depth of anaesthesia monitoring to reduce the risk of under- or overdosing.²⁸ Unfortunately, in our cohort, the model and weight used for the patients were not recorded.

Postoperative care, according to the ERAS guidelines, includes early oral intake, diet progression protocol, vitamin and mineral supplementation, nutritional monitoring, pharmacological and mechanical thromboprophylaxis, and early mobilisation. As our study examined only immediate postoperative complications and outcomes, we did not assess these aspects of postoperative care. The long-term efficacy of the procedure was also not investigated, namely, long-term weight loss, effect on comorbid diseases and long-term complications.

The mortality in this group was almost 7%. In 2024, the IFSO reported overall mortality rates after bariatric surgery ranging from 0% to 0.25% globally. South Africa had a mortality rate of 0.25% (1 in 394 cases).⁹ The risk factors for higher mortality include age above 65 years, BMI > 60 kg/m², severe comorbidities, revisional bariatric surgery and delayed recognition of complications such as anastomotic leak or pulmonary embolism.^29,30 Of the patients who died, high BMI, VTE and anastomotic leak were contributing factors. These deaths emphasise the requirement for vigilant postoperative monitoring, strict thromboprophylaxis and early recognition of surgical complications, which can be challenging in this cohort. The disproportionately high mortality rate in our cohort reinforces the need for strengthened preoperative optimisation and postoperative care measures that must form part of future bariatric surgery programmes.

Several limitations of our study need to be acknowledged. This was a retrospective audit, and the analysis depended on the completeness and accuracy of the records. There may be poor documentation of certain practices and under-reporting of side effects such as PONV. Some eligible patients had to be excluded because no preoperative clinic information was available, potentially introducing selection bias. As this study was conducted as an audit rather than a formal quality improvement project, many relevant stakeholders were not engaged in the review, progression and analysis of the process, thereby limiting multidisciplinary insight. The small sample size reflects the single bariatric unit in which the study was conducted, and caution should be exercised when generalising these results to larger or differently resourced centres. Lastly, compliance with ERAS recommendations was assessed using retrospective data rather than direct observation, and patient-centred postoperative efficacy was not analysed, making it difficult to determine the intervention’s long-term clinical impact.

Future endeavours include establishing an institution-specific ERAS protocol that involves multidisciplinary coordination among anaesthesia, surgery, nursing staff, dietitians, psychology and physiotherapy. This would allow a standardised framework for perioperative management. Once this protocol is implemented, a reaudit should be conducted to assess improvements in compliance and ensure continuity of changes in practice. Lastly, there should be ongoing evaluation analysing the efficacy of the procedure in terms of sustained weight loss, reversal of organ dysfunction and long-term complications in addition to immediate complications and mortality. The incorporation of weight-loss drugs into the optimisation has not yet been delineated. The criteria for bariatric surgery may change over time, considering the advancement of care for obese patients.