Table 3.

Results, outcomes and limitations of included studies

Author. Year (country)	Reported EI (mean ± S.D.) (kcal/day)	TEE (mean ± S.D.) (kcal/day)	Ratio of reported EI:TEE (mean ± S.D. where available), Reporting status	EI mis-reporting results	Limitations
Bandini et al. 1990 (USA) [21]	1935 ± 722 (obese) 2193 ± 618 (non-obese)	3390 ± 613 (obese) 2755 ± 600 (non-obese)	0.59 ± 0.24 (obese), UR0.81 ± 0.19 (non-obese), UR	• Underestimation by: 41.3 ± 23.6% (obese) and 19.4 ± 18.7% (non-obese) (p < 0.0001)• Underestimation greater in obese group compared to non-obese group (p < 0.001)• Mean weight change over study period: 0.37 ± 0.97 kg (obese), 0.08 ± 0.69 kg (non-obese)• Results remained significant after adjustments for body weight changes	• LOA NR• Participants paid to encouraged adherence to FR
Bandini et al. 1999 (USA) [22] (considered as a subset of the above article as the data are from the same study)	1767 ± 601 (obese) 2151 ± 647 (non-obese)	3314 ± 644 (obese) 2799 ± 680 (non-obese)	0.53 (obese), UR0.77 (non-obese), UR	• Underestimation by 44.5 ± 21.8% (obese) and by 21.8 ± 20.5% (non-obese)• Underestimation greater in obese group compared to non-obese group (p < 0.001)	• As above
Börnhorst et al. 2014 (Belgium, Spain and Sweden) [23]	1468 ± 313 (overweight/obese)1511 ± 501 (thin/normal)	1554 ± 214 (overweight/obese)1513 ± 231 (thin/normal)	0.96 ± 0.23 (overweight/obese), AR1.01 ± 0.33 (thin/normal), AR	Underestimation by:• 5.5% in overweight/obese, 86 ± 356 kcal/day• 0.1% in thin/normal, 2 ± 471 kcal/day	• LOA NR• Differences between EI and TEE were large at the individual level• Small, convenience sample• Weight not recorded at end of DLW period
Champagne et al. 1998 (USA) [24]	1707 ± 69 (central fat)1782 ± 70 (obese) 1811 ± 69 (lean) 1883 ± 70 (peripheral fat)	2498 ± 64 (central fat)2338 ± 65 (obese) 2263 ± 64 (lean) 2245 ± 65 (peripheral fat)	0.68 (central fat), UR0.76 (obese), UR0.80 (lean), UR0.84 (peripheral fat), AR	Underestimation by:• 31.5 ± 2.9% in central fat, 791 ± 72 kcal/day• 24.5 ± 3.2% in obese, 557 ± 74 kcal/day• 21.2 ± 3.1% in lean, 453 ± 72 kcal/day• 17.3 ± 3.1% in peripherally fat, 362 ± 74 kcal/day	• LOA NR• Weight NR at end of DLW period• Significance levels reported in paper but unable to be interpreted
Singh et al. 2009 (USA) [25]	2094 ± 563 (Boys) 1890 ± 702 (Girls)	3332 ± 312 (Boys) 2835 ± 336 (Girls)	0.63 ± 0.18 (Boys), UR 0.66 ± 0.22 (Girls), UR	• Reported EI significantly lower than measured TEE (p < 0.0001) regardless of gender• Underestimation by 35 ± 20% in whole group, 1065 ± 636 kcal/day• Underestimation by 34% in girls, 945 ± 626 kcal/day• Underestimation by 37% in boys, 1238 ± 633 kcal/day• Error in reported intake was greater with increasing BMI (p = 0.05)	• LOA NR• EI and TEE not measured over the same time period• Use of convenience population
Vasquez et al. 2006 (Chile) [26]	1846 ± 261 (Boys) 1728 ± 168 (Girls)	1629 ± 183 (Boys) 1481 ± 202 (Girls)	1.13 (Boys), AR 1.17 (Girls), OR	• Overestimation by 13% in boys and 17% in girls (no p value reported)	• LOA NR• Population not representative• Did not report statistical differences between EI and TEE• Change in weight over DLW period NR
Waling and Larsson. 2009 (Sweden) [27]	2182 ± 388 kcal/day	2584 ± 397 kcal/day	0.86 ± 0.16, AR	Underestimation by:• 14% in whole group, 397 ± 421 kcal/day (95% CI: 0.86, 2.45)• 17% in boys, 464 ± 368 kcal/day (95% CI: 0.84, 3.04)• 11% in girls, 335 ± 471 kcal/day (95% CI: 0.07, 2.72)Underestimation did not differ between:• boys and girls (95% CI: −2.17, 1.08)• children who were overweight (n = 16) and those who were obese (n = 5) (95% CI: −1.64, 2.21)three children (14%) reported an EI within ± 5%; two children (10%) reported an EI > 5%; and 16 children (76%) reported an EI < 5% of individually measured TEE	• LOA NR• Small sample size

AR adequate reporters, DLW doubly labelled water, EI energy intake, FR food record, LOA limits of agreement, NR not reported, OR over reporters, TEE total energy expenditure, UR under reporters