Seven randomised controlled trials (RCTs) addressed the impact of oesophageal ultrasound cardiac output monitoring during surgery in comparison to other monitoring options (total n = 583). All included studies were judged to be of a high quality.
Five trials (n=453) assessed the benefits of adding oesophageal Doppler monitoring to CVP plus conventional clinical assessment. No conclusions could be drawn regarding mortality due to the overall low number of deaths. Major complications were reported by three studies (n=220). Overall, there was a significant Peto odds ratio (p=0.00002) indicating a reduction in major complications when oesophageal Doppler monitoring was used. No significant heterogeneity was observed and the sensitivity analyses did not substantially alter the results.
Four trials (n=388) reported total complications of adding oesophageal Doppler monitoring to CVP plus conventional clinical assessment. A random-effects meta-analysis found a significant decrease in total complications when oesophageal Doppler ultrasound monitoring was used, but no summary estimate was presented. All five trials reported length of hospital stay: four found a statistically significant reduction in this outcome when oesophageal Doppler ultrasound monitoring was used, but it was not possible to pool the studies (less than 80 per cent of papers reported usable data).
Two studies (n=130) compared oesophageal Doppler monitoring plus conventional assessment versus conventional assessment alone. Total mortality was not informative due to the small numbers involved. No data was available for major complications. One study reported on total complications and found significantly fewer complications in the Doppler-monitored group, but the percentage of patients with complications did not differ significantly between groups.
Length of hospital stay was assessed in a random-effects meta-analysis, indicating that patients receiving oesophageal Doppler monitoring had a significantly shorter hospital stay (p=0.008) compared to those receiving conventional assessment alone. No effect size was presented due to the low number of studies reporting usable data.
One RCT addressed the impact of oesophageal ultrasound cardiac output monitoring during hospitalisation in comparison to other monitoring options (n=174). The study was judged to be of a high quality and compared oesophageal Doppler monitoring plus CVP and conventional clinical assessment versus CVP plus conventional clinical assessment. No large treatment effects were observed on the outcomes of interest and no conclusions could be drawn.
No studies compared oesophageal Doppler monitoring versus thermodilution with a pulmonary artery catheter.
Adverse effects of monitoring (23 studies: four RCTs; 18 case studies; and one case report). No serious adverse events associated with oesophageal probes were reported in the literature or the MAUDE database. Minor events included incorrect probe placement, tube displacement and minimal trauma to the buccal cavity. Oesophageal Doppler probes can be considered to be relatively low-risk devices.
The original report details all analyses in full including numerous sensitivity analyses which have not been summarised here.