1. Pressure recording analytical method for measuring cardiac output in critically ill children: a validation study 
R. Saxena, A. Durward, N.K. Puppala, I.A. Murdoch, S.M. Tibby 
British Journal of Anaesthesia 
Volume 110, Issue 3, Pages (March 2013)
DOI: /bja/aes420
Copyright © 2013 The Author(s) Terms and Conditions

2. Fig 1
Study protocol. UD, ultrasound dilution; PRAM, pressure recording analytical method. The numbers 1, 2, and 3 refer to the first, second, and third UD measurements, respectively.
British Journal of Anaesthesia  , DOI: ( /bja/aes420)
Copyright © 2013 The Author(s) Terms and Conditions

3. Fig 2
Abnormal UD curves. (A) An inadequate saline indicator injection demonstrated by a double humped blue curve instead of a smooth and narrow curve. Also, the indicator curve (pink) is appearing before the injection curve (blue) terminates. (B) An unsatisfactory UD curve (pink). Two abnormalities are present: (i) a large ‘double hump' seen on the down slope, indicative of significant anatomical shunt; and (ii) an unstable continuation of the indicator curve (green) which moves below the baseline.
British Journal of Anaesthesia  , DOI: ( /bja/aes420)
Copyright © 2013 The Author(s) Terms and Conditions

4. Fig 3
Interpretation of polar plots. (A) A hypothetical change in CO, as measured by two methods. If both methods concur exactly, the vector representing the change will lie on the line of identity. In example 1, the test method is over-reading the increase in CO compared with the reference method by 1 litre min−1. The vector representing this relationship will have an angle θ1 that has a positive deviation from the line of identity. This is represented on the polar plot (B), whereby the angle θ1 is represented relative to the zero degree axis, and the size of the vector is relative to the hypotenuse of the change in CO from the two methods. For a decrease in CO, as seen in example 2, where the test method also overestimates the negative change in CO (θ2), the corresponding polar plot vector will be relative to the 180° axis (B). Good concurrence is represented by 95% of the points lying within an absolute deviation of ±30° from the polar axis (0–180°, blue dashed lines). It is also customary to exclude small changes in CO, identified as those lying within the circular zone of exclusion. Typically, this boundary represents changes in CO that are <10% of the mean change (blue shaded circle). Here example 1 shows adequate concurrence (within the ±30° boundary from the polar axis), whereas example 2 does not.
British Journal of Anaesthesia  , DOI: ( /bja/aes420)
Copyright © 2013 The Author(s) Terms and Conditions

5. Fig 4
Bland–Altman analysis for CO as measured by transpulmonary UD and PRAM.
British Journal of Anaesthesia  , DOI: ( /bja/aes420)
Copyright © 2013 The Author(s) Terms and Conditions

6. Fig 5
Bland–Altman analysis for cardiac index as measured by transpulmonary UD and PRAM.
British Journal of Anaesthesia  , DOI: ( /bja/aes420)
Copyright © 2013 The Author(s) Terms and Conditions

7. Fig 6
Polar plot demonstrating agreement between transpulmonary UD and PRAM for tracking changes in CO. The small, blue-shaded circle indicates the zone of exclusion, whereby changes in CO are too small to assess accurately. Hence data points lying within this circle are disregarded. Elsewhere, good agreement in tracking change in CO is measured by the proportion of data points falling within the polar limits of ±30° (bold, blue dotted lines) from the polar axis.
British Journal of Anaesthesia  , DOI: ( /bja/aes420)
Copyright © 2013 The Author(s) Terms and Conditions