4. Results 
Eligible and included were all scientific publications on interactions of Valeriana officinalis L. root and preparations thereof. Among these are altogether 21 original publications. 11 of them present data from in vitro studies on pharmacokinetic interactions. One of these publications [34] contains, in addition, data on animal studies. Two publications [35, 37] present clinical studies on pharmacokinetic interactions. 8 studies are available on the subject of pharmacodynamic interactions, thereof two in vitro and three animal studies, one clinical study, and in addition two case reports.

4.1. Pharmacokinetic Interactions
Predicting pharmacokinetic herb-drug interactions is difficult because the pharmacological actions of the interacting drugs are often not related. This is also the case with the subject of this review, valerian as a treatment of insomnia, and chemically defined cytostatic therapies. Herb-drug interactions can occur on the levels of absorption, distribution, metabolism, or excretion and can change the amount and duration of the availability of a drug at the site of action. The interactions due to drug metabolism can be, as mentioned above, based on phase 1 metabolism (mainly involving cytochrome P 450 isoenzymes) or, rather rare, on phase 2 metabolism involving, for example, P-glycoproteins (P-gp), which is relevant for outward bound transport processes, for example, in the intestinal wall, or on conjugation, for example, with glucuronic acid [43]. In case that a specific herb-drug interaction is identified, its clinical significance depends on the degree of accumulation and the therapeutic window of the respective drug [44]. Also the dosage, time of administration, galenic properties, and coadministration as well as intrinsic and extrinsic factors may be of importance.

4.2. In Vitro Studies
Budzinski et al. [38] used a fluorometric assay for analysis of the in vitro CYP 3A4 inhibitory capability of dilutions of a valerian fluid extract (no further information available) using human CYP 3A4. An IC50 of 1.8% the undiluted extract in the reaction mixture was identified. The authors stated that “the in vitro interactions, though weak, may have clinical importance.” However, the valerian fluid extract tested was not specified, and the inhibitory concentration identified was very high. Additionally, the use of fluorometric methods is highly susceptible to interference by fluorescent herbal components [45]. Therefore, the results do not allow a valid extrapolation towards a clinically relevant effect.
Lefebvre et al. [39] determined the in vitro effects of 14 commercially available single entity and combination herbal products containing extracts of valerian root, on CYP 3A4-mediated metabolism and on P-gp transport. The extracts were prepared by extracting 100 mg of the powdered commercial preparations with 1 mL of water and 70% ethanol or acetonitrile and characterized by determination of total valerenic acids by HPLC. In a proportion of 1–5% of total assay volume, most extracts showed an inhibition to different extents. Six extracts had some inhibitory effects on P-gp. The authors concluded that “there is wide variation in commercially available samples of valerian root. The findings from this study suggest that valerian root may have an initial inhibitory effect, when taken with therapeutic products. Further work is warranted to determine whether valerian root can affect other CYP 450 isozymes and how the results of this in vitro investigation can be extrapolated to in vivo situations.” Indeed, several of the tested preparations are insufficiently characterized, as only the valerenic acid content is provided. In addition, only one rather high concentration of each extract was tested and, again, a fluorometric method was used, which is highly susceptible to interference by fluorescent herbal components. These limitations do not allow valid extrapolations with respect to a clinical relevance of the results.
The aim of a further study was to evaluate the in vitro effect of commercially available valerian medicinal products on the metabolic activities of the CYP 450 isoenzymes 1A2, 2D6, and 3A4. A valerian extract (drug extraction ratio 6 : 1, extraction solvent ethanol 60%) was incubated with human primary hepatocytes for three times within 48 h. The activities of the CYP isoenzymes were determined by analyzing the metabolites of test substances by HPLC [36, 46]. The herbal extract concentrations used in in vitro metabolic systems were claimed to cover the whole range of herbal concentrations occurring in vivo. Dose dependent and statistically significant increases in CYP 2D6 and CYP 3A4 were observed only with two concentrations (0.188 mg/mL and 1.875 mg/mL). The authors postulate an allosteric antagonism but also clearly point to the limitations of their study, as hepatocytes from only one donor were used. Therefore, the well-known polymorphisms in the CYP superfamily were not covered.
In a follow-up study, the same preparation was used, and an IC50 for CYP 3A4 inhibition at a concentration of 0.756 mg/mL was observed [40]. Slight but significant effects on bidirectional digoxin transport (involving P-gp) were found only with 1.875 mg/mL. The authors conclude that both CYP 3A4 and P-gp interactions are unlikely to be clinically relevant, as the systemic concentrations are probably much lower; therefore, respective IC50 values for P-gp cannot be reached in vivo. This conclusion is also applicable to their previous study [36] and to their study on CYP 2C19, which showed a weak induction by valerian root extracts [47].
In a further in vitro study conducted on mouse and human liver microsomes [34], the postulated inhibitory action of 2.5–75 μg/mL of a valerian extract (aqueous ethanolic, not further defined) on CYP 1A1, CYP 1A2, CYP 2C, and CYP 3A was not confirmed.
In another study, the effect of valerian, valerian/hops extracts, and valerenic acid on the glucuronide conjugation of various substrates (17α-estradiol, acetaminophen, morphine, and testosterone) was determined using human liver microsomes [41]. Test substance was a valerian capsule (250 mg extract corresponding to 3.48 mg valerenic acid per capsule, further specification not available), which was extracted with 80% methanol (5 mL/capsule). Also, the activities of UGT 1A1 and UGT 2B7 were tested in the presence of 2.5 or 5.0 mL of valerian or valerian/hops extract per 250 mL final incubation volume. Valerenic acid significantly inhibited the glucuronidation by both microsomes and UGTs with the rather high concentration of 1 mg/mL. Due to this fact and the very high glucuronidation capacity of the liver, a clinical relevance of these results cannot be assumed.
Mohamed at al. [42] used an UGT 1A1 assay to test a commercially available valerian preparation (≥0.1% valerenic acid; extraction medium 70% ethanol, recommended daily dose 1000 mg) for inhibition of human estradiol-3-O-glucuronidation (E-3-G) in the same concentrations as Hellum et al. [36]. E-3-G was quantified by HPLC. IC50 was 0.562 mg/mL, which would be reached by dissolving one daily dose in a volume of 1.8 L. According to the authors, this volume is in the same order of magnitude as the volume of the intestine, and therefore a potential effect in the intestine could not be excluded [48]. However, even if a transient partial inhibition of UGT 1A1 in the intestine would occur, it would be transient, rather than a long-lasting enzyme induction, and could not to lead to a persistent change of bioavailability of another drug.

4.3. In Vivo Studies
In an in vivo study on mice, a valerian extract (aqueous ethanolic, not further described), given in a dose of 0.5% with the diet over 28 days (corresponding to a daily dose of 595 mg/kg in average) did neither affect the CYP content of the liver nor the activities of CYP 1A1, CYP 1A2, CYP 2C9, and CYP 3A4 [34].
In twelve healthy volunteers (6 males and 6 females, age 30.9 ± 7.2 years, nonsmokers from South Carolina), the effect of a valerian root extract on the activity of the drug-metabolizing enzymes CYP 2D6 and CYP 3A4 was tested [37]. Daily before going to bed, participants took two tablets, with 500 mg valerian extract each (extraction solvent ethanol 70%, valerenic acid content 5.51 mg/tablet), for 14 subsequent days. This dose is well within the range of recommended doses for valerian preparations. The probe drugs dextromethorphan (30 mg; CYP 2D6 activity) and alprazolam (2 mg; CYP 3A4 activity) were administered orally at baseline and after treatment with valerian, and dextromethorphan to dextrorphan metabolic ratios and alprazolam pharmacokinetics were determined. The ratio of dextromethorphan to dextrorphan increased slightly but significantly from 0.214 to 0.254. The maximum concentration of alprazolam was moderately increased for about 20% from 25 to 31 ng/mL (P < 0.05). This change is therefore within the frame of 80% to 125% rated as equivalent by FDA. Bioavailability of other medicines would therefore not be relevantly diminished and therapeutic efficacy of a therapy not questioned. Changes of other pharmacokinetic parameters were not detected. In conclusion valerian in therapeutic doses is unlikely to produce clinically relevant effects on CYP 2D6 and CYP 3A4 pathways which could diminish the therapeutic efficacy of other drugs.
In another study, twelve healthy volunteers (six men and six women, age mean ± SD = 24 ± 3 years, weight 69.3 ± 14.2 kg KGW) were randomly assigned to receive valerian (DER 4 : 1, no standardization claim) for 28 days, three times daily 125 mg [35]. Before and after the test period, the activities of CYP 3A4/5 (1-hydroxymidazolam/midazolam serum ratio), CYP 1A2 (paraxanthine/caffeine serum ratio), CYP 2E1 (hydroxychlorzoxazone/chlorzoxazone serum ratio), and CYP 2D6 (debrisoquine urinary recovery ratio) were determined. All subjects were nonsmokers and extensive metabolizers of CYP 2D6. No changes at all in phenotypic ratios were observed. The daily dose applied in this study was corresponding to 1.5 g drug, which is in the lower range of doses recommended, for example, in the monograph of the HMPC. Despite that, the fact that there was not even a tendency of an effect in the CYP isoenzymes tested underlines the assumption of the lack of an interaction potential in CYP enzymes.

4.4. Pharmacodynamic Interactions
Pharmacodynamic interactions include the concurrent administration of drugs having the same or opposing pharmacologic actions and also the change of the sensitivity or the responsiveness of the tissues to one drug, induced by another one. Many of these interactions can be predicted from knowledge of the pharmacology of each drug. They were proposed for valerian mainly with drugs influencing vigilance such as codeine, citalopram, and benzodiazepines. A presumable interaction with benzodiazepines, which are positive allosteric modulators on GABA-receptors, is based on in vitro data suggesting GABAergic mechanisms of action of valerian extracts [49, 50]. These studies however are inconclusive and require independent replication.
In rats an in vivo study was conducted on interactions between valerian root tincture (ethanol 100%, 1 : 10, not further characterized, daily therapeutic dose 3060 mg for an adult human) and haloperidol with respect to impaired liver or kidney functions. Valerian tincture was applied with the drinking water (1%, corresponding to an extract dose of 200–250 mg/kg bw/d). Haloperidol (38 mg/kg bw) was applied intramuscularly once every 4 weeks over 12 weeks beginning after 15 days of treatment with valerian [51]. While renal effects were lacking, in some of the parameters measured in liver homogenates, slight and statistically significant deviations from control values were observed, suggesting an additive effect of haloperidol and valerian, however of questionable relevance. Authors concluded that in humans a possible toxic additive effect would occur only at supratherapeutic doses. The same working group evaluated the effect of valerian in a rat model of orofacial dyskinesia using the same application procedure as described above. Valerian did not influence orofacial dyskinesia induced by haloperidol. Also oxidative stress parameters were not changed [52].
In another in vivo study on mice [53], 25 mg/kg of a valerian root dry extract was combined with 25 mg/kg of a liquorice extract (both prepared with ethanol 70%, not further characterized) or the benzodiazepine alprazolam (0.7 mg/kg). Tests were conducted in an elevated plus maze. Valerian and alprazolam, rather than liquorice, significantly increased time spent in the open arm, pointing to an anxiolytic effect. The effect of alprazolam, combined with liquorice or valerian, was significantly increased compared to each of the single substances. The authors discuss an improved bioavailability, for example, due to an increased gastrointestinal absorption induced by liquorice. The relevance in humans remains unclear.
In 48 healthy volunteers, pharmacodynamic interactions of single doses of valerian (100 mg/d, extract specification lacking) and propranolol (20 mg/d) were evaluated. The results were not presented in detail but indicated that the two drugs act independently from each other. Interactions with respect to heart rate and parameters of psychic strain could not be demonstrated [54].
A possible pharmacodynamic interaction of valerian preparations with other drugs is supported only by two case reports. A 40-year-old male patient had taken lorazepam (2 mg/d) for two month without side effects. Then he additionally took for two days an infusion of valerian root two hours before bedtime and, just before going to bed, an infusion of valerian root and passion flower herb (dose unknown) without side effects. On the third day he took, instead of the infusion, three tablets containing a combination of valerian root extract (300 mg) and passion flower root and herb extracts (350 mg; no further information available) before bedtime. He thereafter suffered from transient mild handshaking and drowsiness [55]. It was suspected that these symptoms were caused by an interaction between the herbal drugs and lorazepam, as they ceased after stopping herbal treatment. It is however questionable to ascribe the adverse event to valerian alone as the herbal preparation also contained a root and herb extract of passion flower, for which monographs are lacking. As the symptoms are also potential adverse effects of lorazepam itself, and no other similar cases have been described since then, and also a coincidence cannot be ruled out.
Another patient (39-year-old female) had taken for two month a daily dose of two tablets of a St. John's wort preparation and one tablet of a valerian preparation and additionally Loperamide [56]. After this time, she was hospitalized with a severe delirium. She claimed that the herbal treatment was a replacement for the opioid meperidine she had been taken before for reducing migraine. The authors proposed an interaction between the herbal drugs and Loperamide, involving MAO-inhibitory properties. Given that none of these herbal medicines has been proven to have MAO-inhibitory properties, while induction of delirium, and a positive drug screening on opioids was reported, the assumption of an involvement of the valerian preparation does not seem to be plausible.