BACKGROUND Skeletal muscle glycogen is an important energy source for muscle contraction and a key regulator of metabolic responses to exercise. Manipulation of muscle glycogen is therefore a strategy to improve performance in competitions and potentially adaptation to training. However, assessing muscle glycogen in the field is impractical, and there are no normative values for glycogen concentration at rest and during exercise. OBJECTIVE The objective of this study was to meta-analyse the effects of fitness, acute dietary carbohydrate (CHO) availability and other factors on muscle glycogen concentration at rest and during exercise of different durations and intensities. METHODS PubMed was used to search for original articles in English published up until February 2018. Search terms included muscle glycogen and exercise, filtered for humans. The analysis incorporated 181 studies of continuous or intermittent cycling and running by healthy participants, with muscle glycogen at rest and during exercise determined by biochemical analysis of biopsies. METHODS Resting muscle glycogen was determined with a meta-regression mixed model that included fixed effects for fitness status [linear, as maximal oxygen uptake ([Formula: see text]O2max) in mL·kg-1·min-1] and CHO availability (three levels: high, ≥ 6 g·kg-1 of CHO per day for ≥ 3 days or ≥ 7 g·kg-1 CHO per day for ≥ 2 days; low, glycogen depletion and low-CHO diet; and normal, neither high nor low, or not specified in study). Muscle glycogen during exercise was determined with a meta-regression mixed model that included fixed effects for fitness status, resting glycogen [linear, in mmol·kg-1 of dry mass (DM)], exercise duration (five levels, with means of 5, 23, 53 and 116 min, and time to fatigue), and exercise intensity (linear, as percentage of [Formula: see text]O2max); intensity, fitness and resting glycogen were interacted with duration, and there were also fixed effects for exercise modes, CHO ingestion, sex and muscle type. Random effects in both models accounted for between-study variance and within-study repeated measurement. Inferences about differences and changes in glycogen were based on acceptable uncertainty in standardised magnitudes, with thresholds for small, moderate, large and very large of 25, 75, 150 and 250 mmol·kg-1 of DM, respectively. RESULTS The resting glycogen concentration in the vastus lateralis of males with normal CHO availability and [Formula: see text]O2max (mean ± standard deviation, 53 ± 8 mL·kg-1·min-1) was 462 ± 132 mmol·kg-1. High CHO availability was associated with a moderate increase in resting glycogen (102, ± 47 mmol·kg-1; mean ± 90% confidence limits), whereas low availability was associated with a very large decrease (- 253, ± 30 mmol·kg-1). An increase in [Formula: see text]O2max of 10 mL·kg-1·min-1 had small effects with low and normal CHO availability (29, ± 44 and 67, ± 15 mmol·kg-1, respectively) and a moderate effect with high CHO availability (80, ± 40 mmol·kg-1). There were small clear increases in females and the gastrocnemius muscle. Clear modifying effects on glycogen utilisation during exercise were as follows: a 30% [Formula: see text]O2max increase in intensity, small (41, ± 20 mmol·kg-1) at 5 min and moderate (87-134 mmol·kg-1) at all other timepoints; an increase in baseline glycogen of 200 mmol·kg-1, small at 5-23 min (28-59 mmol·kg-1), moderate at 116 min (104, ± 15 mmol·kg-1) and moderate at fatigue (143, ± 33 mmol·kg-1); an increase in [Formula: see text]O2max of 10 mL·kg-1·min-1, mainly clear trivial effects; exercise mode (intermittent vs. continuous) and CHO ingestion, clear trivial effects. Small decreases in utilisation were observed in females (vs. males: - 30, ± 29 mmol·kg-1), gastrocnemius muscle (vs. vastus lateralis: - 31, ± 46 mmol·kg-1) and running (vs. cycling: - 70, ± 32 mmol·kg-1). CONCLUSIONS Dietary CHO availability and fitness are important factors for resting muscle glycogen. Exercise intensity and baseline muscle glycogen are important factors determining glycogen use during exercise, especially with longer exercise duration. The meta-analysed effects may be useful normative values for prescription of endurance exercise.