Description of studies
We identified 387 potentially relevant records, including 11 trial protocols (Online Resource 1). Four protocols corresponded to studies included in our review and 7 lacked published results. One full-text article was excluded due to a language barrier  and 11 studies were included (Fig. 1) .
The studies included 3592 patients (Tables 1 and 2). Two RCTs did not differentiate between acute and chronic cough, although chronic lung diseases such as COPD and asthma were excluded [16, 17]. One RCT included only patients with recurrent acute URTIs (≥ 6/year) . Three OSs included subjects with pneumonia  and chronic respiratory diseases [20, 21].
Risk of bias
Of the RCTs, 2 were found to be at low risk, 3 at high risk, and 1 with some concerns for bias (Fig. 2a). Sources of bias included inadequately described randomization, lack of blinding (2 single-blind [16, 17], 1 open-label ), incomplete baseline data, subjective outcome measurements, and selective reporting of results. Of the non-randomized studies, 4 were found to be at serious and 1 at critical risk of bias (Fig. 2b). Sources of bias included uncontrolled confounders, subjective and unblinded measurement of outcomes, and selective reporting of results. Several RCTs and OSs explicitly allowed concomitant medication for the target condition (expectorants  and antipyretics [18, 22,23,24] or antibiotics [18, 19, 25]).
Overall, most studies were found to be at risk of selection bias due to inadequate descriptions of populations screened for eligibility and selection processes.
Financial conflicts of interest
Six studies declared sponsorship [19, 20, 22,23,24] or were commissioned by  pharmaceutical companies. Five studies did not report funding sources. Of these, 3 were affiliated with the manufacturer of the investigational product [16, 17, 21] and 2 received medications from pharmaceutical companies but did not report affiliations [18, 26].
Effects of interventions
Ivy versus placebo
Two double-blinded RCTs compared an ivy mono-preparation to placebo in 390 adults over 7 days [23, 24] (Table 2). Cough severity was measured by the Bronchitis Severity Scale (BSS) and Visual Analog Scale (VAS) and cough frequency by the Verbal Category Descriptive (VCD) scale (Fig. 3). Statistically significant differences in BSS, VAS, and VCD improvement favoring ivy treatment were reported by treatment day 3.
Two single-blinded RCTs compared an ivy/marsh-mallow/London rocket preparation to placebo in 370 adults and children over 7 days [16, 17]. Physician-measured improvements in cough, congestion, sore throat/chest discomfort, fatigue/weakness, fever, and body ache were reported. A higher percentage of moderate and complete symptom resolution in the intervention group compared to the control was reported.
Ivy versus other therapies
One open-label RCT compared an ivy/thyme preparation to standard care, defined as warm alkaline mineral water, an antipyretic (paracetamol), decongestant drops, and a local antibiotic (fusafungin) in 54 children for 7–10 days . Outcomes included the proportion of patients with wet cough and congestion, number of daily coughing fits, fever, and laboratory markers of inflammation. The difference in coughing fits was statistically significant at treatment days 3–4 (approximately 12 vs. 18 fits/day in treatment versus standard care groups, respectively), but not at subsequent follow-up.
The CCT compared an ivy/marsh-mallow/London rocket preparation to a poly-herbal comparator in 60 adults and children over 15 days . Cough intensity, throat soreness, congestion, sputum production and viscosity, and shortness of breath were measured using an unspecified method. Cough-related quality of life (QoL) was measured via the Leicester Cough Questionnaire (LCQ, Online Resource 2). A statistically significant decrease in all symptoms in the investigational group compared to baseline was reported.
Different ivy formulations
A double-blinded, randomized noninferiority trial including 590 adults and children compared two ivy mono-preparations for 6–8 days . Outcomes included BSS improvement, physician- and patient-/parent-rated efficacy and tolerability, and percent of patients able to return to school or work. The BSS decreased in the entire study population, with a nonsignificant difference between groups.
The four prospective OSs investigated ivy mono-preparations in 2128 patients, 86.2% of which were children [19,20,21, 25]. Three OSs included concomitant antibiotics and cold medications (e.g., decongestants, nasal sprays) in analyses [19, 21, 25]. Of these, one OS included a subgroup analyses; patients who did not receive antibiotics showed slightly higher percentages of clinical worsening .
Adverse events and tolerability
All studies recorded data on adverse events, most of which were gastrointestinal [20,21,22, 24,25,26]. Two mild unspecified allergic reactions were reported  and one isolated skin reaction was possibly related to an ivy mono-preparation . Patient-reported tolerability was reported as good or very good overall [20,21,22, 24, 25].
Heterogeneity of studies
Four studies measured cough severity using the BSS [22,23,24,25] and 1 modified the BSS to include wheezing instead of sputum  (Fig. 3). The heterogeneity of study designs, inclusion criteria, and treatments precluded meta-analysis.