The search yielded 13,033 potential reports. Upon removal of 5549 duplicates, 7484 studies remained for screening based on title and subsequently by abstract. A total of fifteen articles were then screened in full text and assessed for eligibility. The average CONSORT score among all reports found was utilized as a baseline to determine study acceptability. The average score of all reports was 12.65. Six studies were excluded due to a CONSORT score of < 12.5, with the conclusion that nine studies were appropriate for analysis. PRISMA flow chart is presented in Fig. 1.
Assessment of Methodologic Quality and Risk of Bias
The nine studies that were included in the final analysis had a mean quality score of 13. One study demonstrated high selection bias, while eight did not provide adequate information to determine allocation concealment, random sequence generation, and detection bias. All nine studies showed low attrition and reporting bias. Table 1 provides detailed bias assessment for each study.
Study Demographics and Characteristics
Across the nine studies, there were a total of 157 participants, 111 of which were male and 34 were female; one study did not provide sex data. Evaluation of medication types was as follows: one methamphetamine study, three methylphenidate studies, three bupropion studies, one amphetamine study, and one mixed methylphenidate or amphetamine study. There were no studies found evaluating performance effects of atomoxetine, guanfacine, or clonidine that met inclusion criteria. Participant demographics included children with ADHD diagnoses, college students with no indication of level of fitness, average citizens who participated in regular endurance exercise, and trained cyclists. Study design, demographics, and performance measures are reported in Table 2.
Medication-Specific Performance Effects
Study-specific findings are recorded in Table 3. In sum, six reports found significant improvement in athletic performance with use of stimulant medications, with detailed review of respective p values provided below.
Methamphetamine and Amphetamine
The use of an inhaled version of L-methamphetamine did not demonstrate a significant change in distance travelled during cycling trials with use of 16mcg or 48mcg dose (p = 0.81) . Conversely, amphetamine, at a dose of 15 mg per 70 kg bodyweight, was evaluated via cycling and running trials, with findings of increase in acceleration (p < 0.05), knee extension strength (p < 0.01), and time to exhaustion (p < 0.01) .
Performance effects of methylphenidate were tested via badminton skill acquisition , handgrip strength , and cycling timed trials . Study participants provided with 21 mg of methylphenidate were found to have improvement in sportsmanship and effort (p < 0.01) but not in acquisition of sport-specific skills . King et al.  and Roelands et al.  both utilized 20 mg methylphenidate in their studies. The methylphenidate groups demonstrated significantly greater mean force in handgrip strength (p = 0.032) . In cycling trials, methylphenidate groups finished 16% faster (p = 0.049) and had significantly greater power output (p = 0.028) .
One of three studies testing bupropion’s effect on performance found significant improvement in work done in cycling trials (p = 0.042), citing a 7.5 ± 9.6% increase with use of 600 mg of bupropion . Even so, two studies evaluating three different doses of bupropion (150 mg, 300 mg, and 600 mg) found no significant difference in exercise performance by way of time to complete a target amount of work or in maximum power output [22, 23].
One study evaluated physical performance changes among children with established ADHD diagnoses and already on treatment with methylphenidate or amphetamine. An overall increase in work rate was found in participants taking these medications (p < 0.05) .
To better ascertain the clinical significance of each study’s findings, effect sizes were analyzed for each performance measure (Table 4). All medications tested demonstrated an effect on physical performance. No effect was found for two of three bupropion studies, with the third demonstrating small to moderate effect on physical performance. Performance measures associated with the largest positive effect included: exercise performance in a timed trial, power output, and knee extension strength. Methylphenidate had a small to large effect in all performance parameters, whereas amphetamine had predominantly small effect in parameters measured. There was no performance measure that was found to consistently demonstrate small to large effect size among all medications evaluated.
Analysis of Secondary Effects
Through the course of the review of the literature, several secondary effects of medication use were noted to be consistently evaluated, namely in the categories of cardiometabolic effects (i.e. changes in heart rate, blood pressure, oxygen consumption, plasma glucose levels, or plasma lactate levels), core temperature, hormone changes, and ratings of perceived exertion or thermal stress. Table 5 summarizes secondary effects of medication treatment per study. Where the data were available, effect sizes were aggregated and/or calculated, which are subsequently summarized in Table 5.
Seven studies evaluated multiple metabolic factors, including blood pressure, heart rate, VO2max, respiratory exchange ratio, ventilatory equivalent for oxygen, and lactate levels. Heart rate was elevated in the medication groups in four studies, representing methylphenidate (p = 0.046) , bupropion (p = 0.043) , amphetamine (p < 0.001) , and mixed medications (p < 0.05) . Peak VO2 was found to be significantly higher in only one of two studies, in which mixed medications were used (p < 0.05) . No significant differences were identified in all other parameters examined. Of the seven studies, six demonstrated small to moderate effect on all parameters [16, 17, 21,22,23,24], while for one study , the effect sizes could not be calculated with the data provided.
Core temperature was tracked in three studies, all of which found significant increase in core temperature after medication treatment. Specifically, core temperature increased significantly with methylphenidate treatment (p= 0.013) at all time points measured during the study . Similarly, core temperature was higher with bupropion treatment in Cordery et al.  (p = 0.021) and Roelands et al.  (p = 0.030) studies. Two studies provided data to allow for evaluation of effect size and demonstrated moderate to large effect on core temperature with use of bupropion [21, 23].
Four studies examined change in hormone concentration with medication treatment. Three studies found significant increase in hormone concentration after medication treatment, all of which were after treatment with bupropion. Specifically, prolactin (p = 0.043) , cortisol (p < 0.05) and ACTH (p < 0.05), and growth hormone (p = 0.008)  were all increased in the bupropion treatment groups. Only one study provided sufficient data for evaluation of effect size and demonstrated small effect on prolactin increase and moderate effect on FSH increase but no effect on cortisol or LH increase with bupropion use specifically .
Ratings of Perceived Exertion and Thermal Stress
Five studies examined ratings of perceived exertion and/or thermal stress, none of which found any significant differences between placebo and medication treatment groups. Effect sizes for three of these studies were able to be calculated, with finding of large effect with bupropion use in two studies [22, 23]. There was varied effect size in a second study which utilized multiple different stimulant medications .
A meta-analysis was completed for both Roelands et al. [20, 23] studies, as these studies utilized identical experimental designs. Analysis of variance for exercise performance and power output with two-tailed F-test was performed. The null hypothesis was that methylphenidate and bupropion do not impact exercise performance and power output, independently. The calculated F = 1.41 for exercise performance and calculated F = 1.358 for power output were both less than the F statistic of 4.99, indicating that the null hypothesis could not be rejected.
The data were pooled and standardized mean differences were calculated for exercise performance and power output, among the treatment and placebo groups. Forest plots for exercise performance and power output were generated for the two standardized mean differences (Fig. 2). Heterogeneity was minimal for exercise performance (I2 = 30.3, p = 0.2309) but extensive for power output (I2 = 76.9, p = 0.0372), and not explained by any single study.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.