The observed vs. expected lung-to-head ratio does not correct for the effect of gestational age on the lung-to-head ratio

Abstract

Objectives: To assess whether the observed vs. expected lung-to-head ratio (o/e LHR) corrects for the dependence of the LHR on gestational age. Study design: Published data on right lung area (LA) and LHR were used to plot the 50th percentile and different fixed values of the o/e LHR (e.g. 30%) against gestational age from 16–32 weeks. The Z-scores for various fixed o/e LHR values and similar percent value of LA were calculated. The effect of using a fixed LHR or a fixed o/e LHR was tested against gestational age. The o/e LHR-equivalent to a fixed LHR of 1.0 was assessed against gestational age. Results: The LHR and the o/e LHR both increase with gestational age. The Z-score of a given fixed value of the o/e LHR (e.g. 30%) is not similar to the Z-score of the same percent (e.g., 30%) of the expected LA, and thus identifies different proportions of subjects. A fixed o/e LHR (e.g. 30%) results in different populations, depending on the gestational age. The o/e LHR equivalent to an LHR value of 1.0 decreases from 80% at 16 weeks to 30% at 32 weeks. Conclusions: The o/e LHR is not independent of gestational age. Studies using this parameter should be interpreted with caution.

Keywords::

• Congenital diaphragmatic hernia
• lung-to-head ratio
• observed vs. expected lung-to-head ratio
• ultrasound

Declaration of Interest: The authors report no conflicts of interest.

Appendix

1.Calculation of the lung area corresponding to a specific o/e LHR

For a 16 week old fetus, the expected right lung LHR and the standard deviation for this LHR can be calculated using the formulae provided by Peralta et al.:

Expected LHR = −2.2481 + 0.2712 × (gestation in weeks) − 0.0033 × (gestation in weeks)^2

= −2.2481 + 0.2712 × (16) − 0.0033 × (16)^2

= 1.246

SD for LHR= −0.0509 + 0.0178 × (gestation in weeks)

= −0.0509 + 0.0178 × (16)

= 0.234

For an o/e LHR of 60%:

60% Expected LHR = 0.60 × (Expected LHR)

= 0.60 × (1.246)

= 0.7476

SD from Expected LHR = [(60% of Expected LHR) − (Expected LHR)]/(SD for LHR)

= [(1.246) − (0.7476)]/(0.234)

= −2.13

The corresponding right lung LA value is −2.13 SD’s away from expected for a fetus of 16 weeks of gestation:

Expected LA = 815.77 − 152.49 × (gestation in weeks) + 9.0085 × (gestation in weeks)^2 − 0.1305 × (gestation in weeks)^3

= 815.77 − 152.49 × (16) + 9.0085 × (16)^2 − 0.1305 × (16)^3

= 147.6

−2.13 SD LA = −2.13 × [−75.502 + 6.8682 × (gestation in weeks)]

= −2.13 × [−75.502 + 6.8682 × (16)]

= −73.2

Thus the expected LA at −2.13 SD is the expected, 147.6, minus the difference, 73.2, which is 74.4.

Consequently, the equivalent percent of LA (% LA) is given by

% LA = (74.4/147.6) × 100% = 50.4%

2. Analysis of Z-scores for the LHR and LA relative to gestational age

Figure 7a shows the Z-score associated with 60% of the expected LHR (an o/e LHR of 60%) and the Z-score for 60% of the expected LA plotted against gestational age. The graph shows that an o/e LHR of 60% varies with gestational age and identifies different proportions of fetuses depending on gestational age. Similarly, the graph shows that an observed LA of 60% of expected also varies with gestational age and identifies different proportions of fetuses depending on gestational age. The graph demonstrates that even at the same gestational age, a o/e LHR = 60% identifies a different population of fetuses than is identified by the 60% of the lung area. Figure 7b shows the same analysis for ano/e LHR of 25% and 25% of expected LA. Figure 8 shows the proportion of patients that would be identified with an o/e LHR of 60% or with 60% of expected LA. The graph shows that the proportion of patients identified by either method changes over gestational age. Furthermore, at the same gestational age, the proportion of patients identified by both methods is different.