Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

The EBM User's Guide Series is now available at the Canadian Centres for Health Evidence

The questions for diagnostic technology articles are based upon: Keenan SP, Guyatt GH, Sibbald WJ, Cook DJ, Heyland DK, Jaeschke RZ. How to use articles about diagnostic technology: gastric tonometry. Crit Care Med. 1999;27(9):1726-31 [abstract]

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Are transoesophageal Doppler parameters a reliable guide to paediatric haemodynamic status and fluid management?

Tibby SM, Hatherill M, Durward A, Murdoch IA.

Intensive Care Med 2001; 27:201-205. [abstract, full-text for subscribers]

Reviewed by M. Duthie, Specialist Registrar, O. Bagshaw, Consultant ICU, Birmingham Children's Hospital, UK.

Review posted August 18, 2001

I. What is being studied?

Study objective:

Assessment of the reliability of transoesophageal Doppler (TOD) parameters Stroke Distance (SD) [a velocity-time integral which represents the distance travelled by a column of blood during systole] and Corrected Flow Time (FTc) [representing the length of time of systolic blood flow adjusted for heart rate] to

1. correspond to a change in SV following a fluid bolus [SD & FTc compared to CVP}
2. predict an increase in SV of > 10% in response to a 10ml/kg fluid bolus [FTc compared to CVP]
3. monitor SVRI [FTc only]

Study design:

Prospective blinded direct comparison with femoral artery themodilution (FATD). (FATD calculations were done independently from TOD calculations by separate "blinded" clinicians)

The patients investigated:

94 ventilated children: 58 post-cardiac surgery, 29 septicaemia, 5 respiratory disease, 2 Others

Exclusions: Aortic coarctation or arch disease, oesophageal disease, anatomical shunts or valvar regurgitation.

II. Does the technology work as it should?

1. Does the Technology Perform to Specifications in a Laboratory Setting?

For laboratory setting we have substituted previous clinical studies to look at TOD's technical performance.

Yes: Successful use of TOD in Adults (1) and a previous study in children by the authors (2) showed a mean Bias of 0.87% and limits of agreement of +/- 16.8% between SD and SV.

However it should be noted that the TOD equipment used in this study was developed for use in adults and has only been formally assessed in children in the study (2) already mentioned.

The accuracy of the reference method (FATD) had also been previously established by the same group (3) .

Repeatability of both methods in the study was good (median coefficients of variation of 3-4%) with 5 measurements made with each method on all subjects by 2 independent blinded observers.

2. Does the Technology Provide Important Diagnostic Information in a Number of Clinical Situations?

Maybe - A low FTc predicted a positive response to fluid bolus in the children who were not post-cardiac surgery. FTc < 0.394s (an empirically derived value with the best sensitivity [90%], specificity [62%], positive predictive [47%] and negative predictive [94%] values for increase in SV) predicted a 10% increase in SV following 10 ml/kg of either 4.5% Human Albumin, FFP or packed red cells, with an area under receiver operating characteristic curve (ROC) of 0.756 (95% CI 0.584, 0.883). (It is difficult to see why the authors did not round up to a value of FTc < 0.4s which seems much easier to work with than < 0.394s). This compares to an ROC area under the curve of 0.513 (95% CI 0.342,0.683) for CVP in the same cohort and 0.559 (95% CI 0.453, 0.661) for FTc in the whole study group. However this was a sub-group analysis and the authors themselves recommended a further study with a larger cohort of septic patients because of the possible confounding effects of impaired myocardial contractility and low SVRI in sepsis (both can lengthen FTc).

An analysis of the impact of vasopressors and/or vasopressor changes may have been interesting. These effects were also offered as possible reasons for the poor performance of FTc in the post-cardiac surgery cohort of the study group. Considering all patients, the mean bias and precision of percent change in FTc compared to SV was 13% ± 34%.

It is also worth commenting on the validity of a comparison of FTc with CVP as an index of the heart's ability to increase SV following fluid load. Hypovolaemia, for which CVP finds common use is not quite the same concept. The heart may be adequately filled but still capable of increasing SV after a fluid bolus!

3. Does the Technology Provide Information That Allows a More Accurate Assessment of the Presence or Severity of Disease in Patients?

Yes - TOD derived SD performed well compared to SV derived by FATD in all patients, giving a mean bias of 1.8% (representing the difference between percentage change following fluid bolus in the test variable SD and the reference variable SV). However the agreement limits of +/-17% are a little wide and the authors chose to omit the Bland-Altman plot showing them.

Importantly SD is continuously monitored, and so it can provide improved warning of states of reduced SV compared to intermittent FATD measurement. By contrast FTc correlated poorly with FATD derived SVRI. This may be highlighting a deficiency in the reference method's ability to accurately measure reductions in SVRI, possibly due to differential effects of sepsis on pulmonary and systemic vascular beds.

III. What Is The Impact Of The Diagnostic Technology?

1. Does the Technology Increase Healthcare Worker Confidence?

Potentially Yes - The main current method for pre-load adjustment is CVP which was consistently out-performed by Stroke Distance (with the distinction between these two noted above). Although the question of healthcare worker confidence is difficult to assess in the early phases of use of a technology, TOD's continuous monitoring ability should also prove useful in this regard to supplement the indirect measures of blood pressure, urine output, etc. now in use for assessing the effects of numerous interventions.

2. Are Therapeutic Decisions Altered as a Result of the Technology?

Yes - The administration of fluid was guided by it in this study (and potentially, if it proves robust in other patients and other studies it could play a significant therapeutic role).

3. Does Application of the Technology Result in a Benefit to the Patients?

Not assessed - Outcome measures were not examined and all the subjects in this study had TOD monitoring (self-controls).

IV. Can I Apply The Diagnostic Technology In My Practice?

1. Can I Expect a Similar Benefit in My Setting?

See III 3.- Anecdotally the experience of one of the reviewers (O. Bagshaw) is that TOD has helped optimise fluid and inotrope usage in several children.

2. Are the Expected Benefits Worth the Associated Costs?

Not assessed - As benefit was not assessed, a cost/benefit analysis will also await further study. If proved to be accurate and reliable over time, this technology has the potential to significantly decrease the risks of invasive thermodilution catheters, the costs of repeated ECHO's, and provide important continual monitoring.

References

1. Singer M, Bennett ED. Noninvasive optimization of left ventricular filling using esophageal Doppler. Crit Care Med. 1991;19(9):1132-7. [abstract]
2. Tibby SM et al.Use of transesophageal Doppler ultrasonography in ventilated pediatric patients: derivation of cardiac output. Crit Care Med. 2000;28(6):2045-50. [abstract]
3. Tibby SM et al.Clinical validation of cardiac output measurements using femoral artery thermodilution with direct Fick in ventilated children and infants. Intensive Care Med. 1997;23(9):987-91. [abstract]

 

Comments

Back to the EB Journal Club Index