Table 1.

Summary of findings.

Outcome & subgroupa	Participants(studies)	SMD [95% CI]	P-value (Z test)	I2 (%)	P-value (Chi2 test)	Interpretation	Certainty of evidence (GRADE rating)
Cortisol, resting
All studies	245 (22)	0.18 [−0.06, 0.42]	0.15	32	0.08
Long-term LC diets (MP-LC diets only)	90 (7)	−0.28 [−0.7, 0.15]	0.2	36	0.15	Long-term, MP-LC vs HC diets neither increase nor decrease resting cortisol.	Moderate ⊕⊕⊕⊝ Downgraded due to unexplained heterogeneity
Short-term LC diets	155 (15)	0.41 [0.16, 0.66]	<0.01	0	0.63	Short-term LC vs HC diets moderately increase resting cortisol.	High ⊕⊕⊕⊕
Test for subgroup differences				86.7	<0.01
Cortisol, 0 h post-exercise
All studies	129 (12)	0.58 [0.17, 0.99]	<0.01	50	0.02
Long-duration exercise	112 (10)	0.78 [0.47, 1.1]	<0.01	0	0.51	LC vs HC diets result in much higher 0 h post-exercise cortisol, after long-duration exercise.	Moderate ⊕⊕⊕⊝ Downgraded due to indirectness (only includes two long-term LC diet studies)b
Short-duration exercise	17 (2)	−0.67 [−1.37, 0.03]	0.06	0	0.78	LC vs HC diets result in much lower 0 h post-exercise cortisol, after short-duration exercise.	Low ⊕⊕⊝⊝ Downgraded due to imprecision (small sample size and p > 0.05)
Test for subgroup differences				92.8	<0.01
Cortisol, 1 h post-exercise
All studies (long-duration exercise only)	55 (5)	0.81 [0.31, 1.31]	<0.01	24	0.26	LC vs HC diets result in much higher 1 h post-exercise cortisol, after long-duration exercise.	Moderate ⊕⊕⊕⊝ Downgraded due to indirectness (only includes one long-term LC diet study)b
Cortisol, 2 h post-exercise
All studies (short-term, MP-LC diets and long-duration exercise only)	36 (3)	0.82 [0.33, 1.3]	<0.01	0	0.79	Short-term, MP-LC vs HC diets result in much higher 2 h post-exercise cortisol, after long-duration exercise.	Moderate ⊕⊕⊕⊝ Downgraded due imprecision (small sample size)
Total testosterone, resting
All studies	155 (13)	−0.48 [−0.87, −0.09]	0.01	49	0.02
MP-LC diets	129 (10)	−0.31 [−0.74, 0.13]	0.17	48	0.04	MP-LC vs HC diets neither increase nor decrease resting total testosterone.	Low ⊕⊕⊝⊝ Downgraded due to indirectness (only includes two short-term LC diet studies)b and risk of bias (randomized studies only: SMD = −0.79, p < 0.01, I2 = 0%)
HP-LC diets (short-term LC diets only)	26 (3)	−1.08 [−1.67, −0.48]	<0.01	0	0.85	Short-term HP-LC vs HC diets greatly decrease resting total testosterone.	Moderate ⊕⊕⊕⊝ Downgraded due to imprecision (small sample size)
Test for subgroup differences				76.5	0.04
Total testosterone, 0 h post-exercise
All studies	28 (3)	−0.03 [−0.95, 0.89]	0.95	65	0.06
Long-term LC diets (MP-LC diets only)	19 (2)	0.44 [−0.21, 1.09]	0.18	0	0.9	Long-term, MP-LC vs HC diets result in higher 0 h post-exercise total testosterone.	Very low ⊕⊝⊝⊝ Downgraded due to indirectness (only includes one long- and one short-duration exercise study)c and imprecision (small sample size and p > 0.05)
Short-term LC diets (HP-LC diets and long-duration exercise only)	9 (1)	−1.01 [−2, −0.01]	0.05	NA	NA	Short-term, HP-LC vs HC diets result in much lower 0 h post-exercise total testosterone, after long-duration exercise.	Low ⊕⊕⊝⊝ Downgraded due to imprecision (small sample size and p = 0.05)
Test for subgroup differences				82.5	0.02

^a Long-term (≥3 weeks), short-term (<3 weeks), long-duration exercise (≥20 min), short-duration exercise (<20 min), MP (<35% protein), HP (≥35% protein).

^b Indirect evidence drawn from long or short LC diets, to support a conclusion about LC diets in general.

^c Indirect evidence drawn from long- or short-duration exercise, to support a conclusion about exercise in general.

CI: confidence interval; HC: high-carbohydrate; HP: high-protein; LC: low-carbohydrate; MP: moderate-protein; SMD: standardized mean difference.