Comparison of clinical characteristic and biochemical marker models in data imputed with original biochemical markers or natural log-transformed biochemical markers

John Allotey; Kym IE Snell; Melanie Smuk; Richard Hooper; Claire L Chan; Asif Ahmed; Lucy C Chappell; Peter von Dadelszen; Julie Dodds; Marcus Green; Louise Kenny; Asma Khalil; Khalid S Khan; Ben W Mol; Jenny Myers; Lucilla Poston; Basky Thilaganathan; Anne C Staff; Gordon CS Smith; Wessel Ganzevoort; Hannele Laivuori; Anthony O Odibo; Javier A Ramírez; John Kingdom; George Daskalakis; Diane Farrar; Ahmet A Baschat; Paul T Seed; Federico Prefumo; Fabricio da Silva Costa; Henk Groen; Francois Audibert; Jacques Masse; Ragnhild B Skråstad; Kjell Å Salvesen; Camilla Haavaldsen; Chie Nagata; Alice R Rumbold; Seppo Heinonen; Lisa M Askie; Luc JM Smits; Christina A Vinter; Per M Magnus; Kajantie Eero; Pia M Villa; Anne K Jenum; Louise B Andersen; Jane E Norman; Akihide Ohkuchi; Anne Eskild; Sohinee Bhattacharya; Fionnuala M McAuliffe; Alberto Galindo; Ignacio Herraiz; Lionel Carbillon; Kerstin Klipstein-Grobusch; SeonAe Yeo; Helena J Teede; Joyce L Browne; Karel GM Moons; Richard D Riley; Shakila Thangaratinam

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Allotey J, Snell KIE, Smuk M, et al. Validation and development of models using clinical, biochemical and ultrasound markers for predicting pre-eclampsia: an individual participant data meta-analysis. Southampton (UK): NIHR Journals Library; 2020 Dec. (Health Technology Assessment, No. 24.72.)

Cover of Validation and development of models using clinical, biochemical and ultrasound markers for predicting pre-eclampsia: an individual participant data meta-analysis

Validation and development of models using clinical, biochemical and ultrasound markers for predicting pre-eclampsia: an individual participant data meta-analysis.

Show details

Contents

< Prev Next >

Appendix 18Comparison of clinical characteristic and biochemical marker models in data imputed with original biochemical markers or natural log-transformed biochemical markers

TABLE 41

First-trimester models for any pre-eclampsia

Variable	Original biochemical markers			Ln(biochemical markers)
Variable	Beta coefficient (95% CI)	OR (95% CI)	p-value	Beta coefficient (95% CI)	OR (95% CI)	p-value
Age (years)	–0.023 (–0.038 to –0.008)	0.977 (0.962 to 0.992)	0.003	–0.023 (–0.039 to –0.007)	0.977 (0.962 to 0.993)	0.006
SBP	0.016 (0.004 to 0.028)	1.016 (1.004 to 1.028)	0.011	0.024 (0.014 to 0.034)	1.024 (1.014 to 1.034)	< 0.001
BMI	0.041 (0.009 to 0.074)	1.042 (1.009 to 1.077)	0.015	0.036 (0.016 to 0.056)	1.037 (1.016 to 1.057)	< 0.001
Nulliparous	1.136 (0.597 to 1.675)	3.113 (1.816 to 5.337)	< 0.001	1.267 (0.703 to 1.831)	3.551 (2.020 to 6.241)	< 0.001
Previous PE	1.350 (0.812 to 1.888)	3.857 (2.252 to 6.607)	< 0.001	1.488 (0.901 to 2.076)	4.430 (2.461 to 7.973)	< 0.001
Renal disease	0.957 (0.319 to 1.594)	2.603 (1.376 to 4.925)	0.004	1.058 (0.449 to 1.667)	2.880 (1.566 to 5.296)	0.001
Hypertension	2.241 (1.962 to 2.520)	9.403 (7.113 to 12.430)	< 0.001	2.215 (1.949 to 2.482)	9.165 (7.024 to 11.960)	< 0.001
Diabetes	0.021 (–0.915 to 0.957)	1.021 (0.401 to 2.605)	0.965	0.178 (–0.737 to 1.094)	1.195 (0.479 to 2.985)	0.702
sFlt-1	–2.42 × 10^–4 (–4.3 × 10^–4 to –5.5 × 10^–5)	0.9998 (0.9996 to 0.9999)	0.012	–0.052 (–0.109 to 0.005)	0.949 (0.897 to 1.005)	0.071
PlGF	–0.005 (–0.008 to –0.002)	0.995 (0.992 to 0.998)	0.001	–0.030 (–0.062 to 0.002)	0.970 (0.940 to 1.002)	0.064
Intercept	–5.893 (–7.478 to –4.309)			–7.022 (–8.515 to –5.530)

TABLE 42

Performance statistics for first-trimester models for any pre-eclampsia

Model	C-statistic			Calibration slope			Calibration-in-the-large
Model	Estimate (95% CI)	τ² (for logit C-statistic)	I² (for logit C-statistic)	Estimate (95% CI)	τ²	I²	Estimate (95% CI)	τ²	I²
Original biochemical markers	0.696 (0.516 to 0.832)	0.205	90.7	0.901 (0.355 to 1.447)	0.094	86.7	–0.004 (–0.939 to 0.931)	0.333	97.9
Ln(biochemical markers)	0.693 (0.538 to 0.814)	0.150	88.8	0.896 (0.302 to 1.491)	0.120	90.5	0.008 (–0.874 to 0.889)	0.291	97.1

TABLE 43

First-trimester models for early pre-eclampsia

Variable	Original biochemical markers			Ln(biochemical markers)
Variable	Beta coefficient (95% CI)	OR (95% CI)	p-value	Beta coefficient (95% CI)	OR (95% CI)	p-value
Age (years)	–0.005 (–0.047 to 0.038)	0.995 (0.954 to 1.039)	0.831	–0.013 (–0.054 to 0.028)	0.987 (0.947 to 1.028)	0.523
SBP	0.053 (0.025 to 0.082)	1.055 (1.026 to 1.085)	< 0.001	0.024 (–0.007 to 0.054)	1.024 (0.993 to 1.055)	0.125
DBP	–0.004 (–0.04 to 0.033)	0.996 (0.960 to 1.033)	0.836	0.042 (0.004 to 0.080)	1.043 (1.004 to 1.083)	0.031
BMI	–0.007 (–0.069 to 0.056)	0.993 (0.933 to 1.057)	0.831	0.006 (–0.049 to 0.061)	1.006 (0.952 to 1.062)	0.837
Previous PE	1.488 (0.694 to 2.281)	4.427 (2.002 to 9.789)	< 0.001	1.689 (0.909 to 2.469)	5.412 (2.481 to 11.805)	< 0.001
Renal disease	1.481 (0.401 to 2.560)	4.397 (1.494 to 12.942)	0.007	1.414 (0.325 to 2.503)	4.112 (1.384 to 12.216)	0.011
Hypertension	0.833 (0.033 to 1.633)	2.300 (1.034 to 5.117)	0.041	0.978 (0.201 to 1.755)	2.659 (1.222 to 5.786)	0.014
sFlt-1	–5.15 × 10^–4 (–1.1 × 10^–3 to 6.1 × 10^–5)	0.9995 (0.9989 to 1.0001)	0.079
Intercept	–10.206 (–13.561 to –6.851)			–10.929 (–13.875 to –7.983)

TABLE 44

Performance statistics for first-trimester models for early pre-eclampsia

Model	C-statistic			Calibration slope			Calibration-in-the-large
Model	Estimate (95% CI)	τ² (for logit C-statistic)	I² (for logit C-statistic)	Estimate (95% CI)	τ²	I²	Estimate (95% CI)	τ²	I²
Original biochemical markers	0.762 (0.576 to 0.883)	0.107	37.0	1.038 (0.666 to 1.410)	0	0	0.036 (–1.269 to 1.342)	0.610	91.7
Ln(biochemical markers)	0.745 (0.591 to 0.855)	0.053	27.2	1.022 (0.695 to 1.350)	0	0	0.014 (–1.240 to 1.268)	0.552	90.2

TABLE 45

First-trimester models for late pre-eclampsia

Variable	Original biochemical markers			Ln(biochemical markers)
Variable	Beta coefficient (95% CI)	OR (95% CI)	p-value	Beta coefficient (95% CI)	OR (95% CI)	p-value
Age (years)	–0.025 (–0.040 to –0.009)	0.976 (0.961 to 0.991)	0.002	–0.024 (–0.042 to –0.006)	0.976 (0.959 to 0.994)	0.009
SBP	0.009 (–0.002 to 0.020)	1.009 (0.998 to 1.020)	0.113	0.020 (0.010 to 0.030)	1.020 (1.010 to 1.031)	< 0.001
BMI	0.051 (0.024 to 0.077)	1.052 (1.024 to 1.080)	< 0.001	0.040 (0.018 to 0.061)	1.040 (1.019 to 1.063)	< 0.001
Nulliparous	1.409 (0.814 to 2.004)	4.093 (2.258 to 7.418)	< 0.001	1.683 (1.035 to 2.330)	5.379 (2.816 to 10.276)	< 0.001
Previous PE	1.197 (0.557 to 1.838)	3.311 (1.745 to 6.283)	< 0.001	1.456 (0.774 to 2.138)	4.289 (2.168 to 8.486)	< 0.001
Renal disease	0.758 (0.084 to 1.433)	2.135 (1.088 to 4.190)	0.028	0.886 (0.225 to 1.547)	2.426 (1.253 to 4.698)	0.009
Hypertension	2.374 (2.084 to 2.663)	10.739 (8.039 to 14.346)	< 0.001	2.326 (2.049 to 2.604)	10.238 (7.756 to 13.515)	< 0.001
sFlt-1	–2 × 10^–4 (–3.8 × 10^–4 to –2.4 × 10^–5)	0.9998 (0.9996 to 1.0000)	0.026	–0.06 (–0.119 to –0.001)	0.942 (0.888 to 0.999)	0.048
PlGF	–0.006 (–0.009 to –0.003)	0.994 (0.991 to 0.997)	0.001	–0.032 (–0.066 to 0.003)	0.969 (0.936 to 1.003)	0.072
Intercept	–5.787 (–7.28 to –4.293)			–7.168 (–8.776 to –5.559)

TABLE 46

Performance statistics for first-trimester models for late pre-eclampsia

Model	C-statistic			Calibration slope			Calibration-in-the-large
Model	Estimate (95% CI)	τ² (for logit C-statistic)	I² (for logit C-statistic)	Estimate (95% CI)	τ²	I²	Estimate (95% CI)	τ²	I²
Original biochemical markers	0.688 (0.505 to 0.827)	0.212	91.9	0.857 (0.338 to 1.376)	0.085	84.4	–0.005 (–0.932 to 0.922)	0.322	97.5
Ln(biochemical markers)	0.685 (0.511 to 0.818)	0.188	90.8	0.849 (0.277 to 1.422)	0.112	88.6	0.010 (–0.856 to 0.876)	0.276	96.5

TABLE 47

Second-trimester models for any pre-eclampsia

Variable	Original biochemical markers			Ln(biochemical markers)
Variable	Beta coefficient (95% CI)	OR (95% CI)	p-value	Beta coefficient (95% CI)	OR (95% CI)	p-value
Age (years)	–0.019 (–0.034 to –0.004)	0.981 (0.966 to 0.996)	0.012	–0.026 (–0.041 to –0.011)	0.974 (0.960 to 0.989)	0.001
SBP	0.019 (0.009 to 0.029)	1.019 (1.009 to 1.030)	< 0.001	0.005 (–0.005 to 0.015)	1.005 (0.995 to 1.015)	0.357
DBP	–0.009 (–0.021 to 0.004)	0.991 (0.979 to 1.004)	0.167	0.020 (0.006 to 0.033)	1.020 (1.006 to 1.034)	0.004
BMI	0.069 (0.055 to 0.084)	1.072 (1.057 to 1.088)	< 0.001	0.069 (0.055 to 0.084)	1.072 (1.057 to 1.087)	< 0.001
Nulliparous	2.099 (1.263 to 2.935)	8.156 (3.534 to 18.820)	< 0.001	2.144 (1.344 to 2.944)	8.532 (3.833 to 18.993)	< 0.001
Previous PE	1.831 (1.013 to 2.648)	6.237 (2.755 to 14.123)	< 0.001	1.947 (1.172 to 2.722)	7.006 (3.227 to 15.209)	< 0.001
Renal disease	1.423 (0.609 to 2.237)	4.150 (1.838 to 9.368)	0.001	1.467 (0.658 to 2.275)	4.335 (1.931 to 9.732)	< 0.001
Hypertension	2.401 (2.120 to 2.682)	11.034 (8.334 to 14.608)	< 0.001	2.389 (2.112 to 2.666)	10.905 (8.268 to 14.383)	< 0.001
PlGF	–0.003 (–0.004 to –0.002)	0.997 (0.996 to 0.998)	< 0.001	–0.539 (–0.658 to –0.420)	0.583 (0.518 to 0.657)	< 0.001
Intercept	–7.895 (–9.314 to –6.476)			–6.070 (–7.559 to –4.581)

TABLE 48

Performance statistics for second-trimester models for any pre-eclampsia

Model	C-statistic			Calibration slope			Calibration-in-the-large
Model	Estimate (95% CI)	τ² (for logit C-statistic)	I² (for logit C-statistic)	Estimate (95% CI)	τ²	I²	Estimate (95% CI)	τ²	I²
Original biochemical markers	0.754 (0.526 to 0.894)	0.151	92.4	0.961 (0.358 to 1.564)	0.052	88.7	0.004 (–1.284 to 1.293)	0.262	98.0
Ln(biochemical markers)	0.746 (0.521 to 0.888)	0.150	93.6	0.945 (0.218 to 1.672)	0.081	94.1	0.001 (–1.859 to 1.862)	0.556	99.0

TABLE 49

Second-trimester models for early pre-eclampsia

Variable	Original biochemical markers			Ln(biochemical markers)
Variable	Beta coefficient (95% CI)	OR (95% CI)	p-value	Beta coefficient (95% CI)	OR (95% CI)	p-value
DBP	0.071 (0.044 to 0.099)	1.074 (1.045 to 1.104)	< 0.001	0.065 (0.039 to 0.092)	1.068 (1.039 to 1.096)	< 0.001
Previous PE	1.998 (1.046 to 2.949)	7.371 (2.847 to 19.084)	< 0.001	2.360 (1.529 to 3.191)	10.594 (4.615 to 24.318)	< 0.001
Renal disease	2.291 (0.151 to 4.431)	9.883 (1.163 to 83.975)	0.036	3.148 (0.961 to 5.335)	23.285 (2.613 to 207.474)	0.005
Autoimmune disease	1.012 (–1.075 to 3.100)	2.752 (0.341 to 22.193)	0.342	1.154 (–0.924 to 3.231)	3.169 (0.397 to 25.315)	0.277
sFlt-1				0.458 (0.096 to 0.820)	1.581 (1.100 to 2.271)	0.013
PlGF	–0.003 (–0.005 to –0.001)	0.997 (0.995 to 0.999)	0.010	–0.988 (–1.283 to –0.693)	0.372 (0.277 to 0.500)	< 0.001
Intercept	–10.116 (–12.119 to –8.114)			–9.343 (–12.661 to –6.025)

TABLE 50

Performance statistics for second-trimester models for early pre-eclampsia

Model	C-statistic			Calibration slope			Calibration-in-the-large
Model	Estimate (95% CI)	τ² (for logit C-statistic)	I² (for logit C-statistic)	Estimate (95% CI)	τ²	I²	Estimate (95% CI)	τ²	I²
Original biochemical markers	0.786 (0.389 to 0.955)	0.304	61.2	1.054 (0.475 to 1.633)	0	0	0.024 (–1.308 to 1.357)	0.229	80.9
Ln(biochemical markers)	0.830 (0.629 to 0.934)	0.041	17.4	1.079 (0.567 to 1.591)	0.011	26.0	0.099 (–0.972 to 1.169)	0.107	66.4

TABLE 51

Second-trimester models for late pre-eclampsia

Variables	Original biochemical markers			Ln(biochemical markers)
Variables	Beta coefficient (95% CI)	OR (95% CI)	p-value	Beta coefficient (95% CI)	OR (95% CI)	p-value
Age (years)	–0.020 (–0.036 to –0.004)	0.980 (0.965 to 0.996)	0.012	–0.027 (–0.043 to –0.012)	0.973 (0.958 to 0.989)	0.001
SBP	0.018 (0.008 to 0.029)	1.018 (1.008 to 1.029)	0.001	0.002 (–0.009 to 0.014)	1.002 (0.991 to 1.014)	0.697
DBP	–0.014 (–0.028 to –0.001)	0.986 (0.973 to 0.999)	0.031	0.018 (0.002 to 0.033)	1.018 (1.002 to 1.034)	0.026
BMI	0.073 (0.058 to 0.088)	1.076 (1.060 to 1.093)	< 0.001	0.074 (0.059 to 0.089)	1.077 (1.061 to 1.093)	< 0.001
Nulliparous	2.539 (1.564 to 3.515)	12.673 (4.777 to 33.622)	< 0.001	2.567 (1.647 to 3.487)	13.023 (5.190 to 32.681)	< 0.001
Previous PE	1.734 (0.799 to 2.669)	5.664 (2.223 to 14.431)	< 0.001	1.839 (0.962 to 2.716)	6.291 (2.617 to 15.120)	< 0.001
Renal disease	1.303 (0.463 to 2.143)	3.680 (1.588 to 8.524)	0.002	1.318 (0.487 to 2.148)	3.735 (1.628 to 8.572)	0.002
Hypertension	2.432 (2.144 to 2.720)	11.380 (8.531 to 15.180)	< 0.001	2.426 (2.144 to 2.709)	11.317 (8.531 to 15.012)	< 0.001
PlGF	–0.003 (–0.004 to –0.002)	0.997 (0.996 to 0.998)	< 0.001	–0.424 (–0.545 to –0.302)	0.655 (0.580 to 0.739)	< 0.001
Intercept	–8.054 (–9.598 to –6.510)			–6.760 (–8.323 to –5.198)

TABLE 52

Performance statistics for second-trimester models for late pre-eclampsia

Model	C-statistic			Calibration slope			Calibration-in-the-large
Model	Estimate (95% CI)	τ² (for logit C-statistic)	I² (for logit C-statistic)	Estimate (95% CI)	τ²	I²	Estimate (95% CI)	τ²	I²
Original biochemical markers	0.746 (0.493 to 0.898)	0.181	93.2	0.936 (0.355 to 1.517)	0.047	86.5	0.004 (–1.295 to 1.303)	0.266	97.8
Ln(biochemical markers)	0.730 (0.477 to 0.889)	0.180	94.4	0.909 (0.186 to 1.632)	0.078	92.8	0.001 (–1.697 to 1.700)	0.461	98.7

Copyright © Queen’s Printer and Controller of HMSO 2020. This work was produced by Allotey et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Bookshelf ID: NBK565561

Contents

< Prev Next >

PubReader
Print View
Cite this Page
Allotey J, Snell KIE, Smuk M, et al. Validation and development of models using clinical, biochemical and ultrasound markers for predicting pre-eclampsia: an individual participant data meta-analysis. Southampton (UK): NIHR Journals Library; 2020 Dec. (Health Technology Assessment, No. 24.72.) Appendix 18, Comparison of clinical characteristic and biochemical marker models in data imputed with original biochemical markers or natural log-transformed biochemical markers.
PDF version of this title (15M)

Other titles in this collection

Health Technology Assessment

Recent Activity

Clear Turn Off Turn On

Comparison of clinical characteristic and biochemical marker models in data impu...
Comparison of clinical characteristic and biochemical marker models in data imputed with original biochemical markers or natural log-transformed biochemical markers - Validation and development of models using clinical, biochemical and ultrasound markers for predicting pre-eclampsia: an individual participant data meta-analysis
Number and percentage of children failing according to each component of the com...
Number and percentage of children failing according to each component of the composite end point - Adenoidectomy with or without grommets for children with otitis media: an individual patient data meta-analysis
End of trial for pump participants: standard operating procedure - A cluster ran...
End of trial for pump participants: standard operating procedure - A cluster randomised trial, cost-effectiveness analysis and psychosocial evaluation of insulin pump therapy compared with multiple injections during flexible intensive insulin therapy for type 1 diabetes: the REPOSE Trial
Calculating utility scores from van Duijvendijk and colleagues - A systematic re...
Calculating utility scores from van Duijvendijk and colleagues - A systematic review and economic evaluation of diagnostic strategies for Lynch syndrome
Additional health economic evaluation information - The Effectiveness, cost-effe...
Additional health economic evaluation information - The Effectiveness, cost-effectiveness and acceptability of Community versus Hospital Eye Service follow-up for patients with neovascular age-related macular degeneration with quiescent disease (ECHoES): a virtual randomised balanced incomplete block trial

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

Bookshelf