Twenty-four RCTs, with 1,285 participants, were included in the review. Sixteen were crossover trials and 8 were parallel-group trials.
The mean Jadad score was 5.9 (standard deviation 1.7) out of a maximum possible score of 8. The scores ranged from 1 to 8. Many of the included studies pooled baseline vital sign data for the treatment and placebo groups.
Pseudoephedrine was found to cause a statistically significant increase in systolic BP (WMD 0.99 mmHg, 95% CI: 0.08, 1.90) and HR (2.83 beats/minute, 95% CI: 2.03, 3.63) compared with placebo. The increase in diastolic BP was not statistically significant (0.63 mmHg, 95% CI: -0.10, 1.35). No clinically significant adverse events were reported.
There was no evidence of publication bias for either systolic or diastolic BP, but there was for HR. With trim-and-fill, the effect was reduced to 1.7 beats/minute (95% CI: 0.96, 2.6).
All effectiveness results were heterogeneous (P<0.001). The subgroup analyses found that while immediate-release formulations significantly elevated systolic BP, sustained release formulations had no effect. Formulation type did not alter the overall effect on diastolic BP. HR was significantly elevated with immediate- and sustained-release formulations. In participants with known, stable (treated) hypertension, the results were the same as overall, but the increase in HR was not statistically significant. There were no significant effects in participants who underwent exercise testing. However, the authors warned that these comparisons might have been underpowered.
There were no effects of age, country or year of study, study design, or duration of washout on the results. No single study overly affected the results. The treatment effects on systolic BP and HR were normally distributed. A dose-response relationship was found for all outcomes. Longer study duration was associated with less of an effect on systolic BP, but had no effect on diastolic BP or HR. Studies with more women demonstrated less effect on all three outcomes. The authors stated that separate analyses of studies with separate baseline data for the placebo and control groups and higher quality studies showed less pronounced effects on outcome, though this analysis was not reported.