Numerous herbal medicines have effects in the central nervous system and have the potential for adversely affecting patients with epilepsy [1
]. New-onset seizures have been reported to occur in patients taking CAM when no other cause has been found [2
]. We report a case of status epilepticus associated with the ingestion of borage oil. Borage (B. officinalis
) oil has been used for the treatment of depression, inflammation, fevers, and coughs, although these uses have not been empirically tested. GLA, a polyunsaturated fatty acid, is the ingredient of borage seed oil that is reported to be responsible for its beneficial, specifically anti-inflammatory health effects [7
]. Borage oil products typically contain 20–27% GLA [7
Borage seed oil is a significant source of gamma-linolenic acid, C18:3(n-6), which is converted to dihomo-gamma-linolenic acid (DGLA) a precursor to a variety of the 1-series prostaglandins and 3-series leukotrienes. GLA is thought to exert its therapeutic effect in rheumatologic and dermatologic conditions by inhibiting leukotriene synthesis, resulting in anti-inflammatory and anti-thrombotic effects [7
]. Borage seed oil has been estimated to contain the highest amounts of GLA: 23% compared to black currant seed oil or evening primrose oil, both of which are also well known herbal supplement sources of GLA [7
]. Borage seed oil also contains high concentrations (up to 35–38%) of LA, another omega-6 fatty acid, and a precursor to GLA in normal fatty acid metabolism. Two putative toxic minor metabolites of LA, leukotoxindiol (LTX), and isoleukotoxindiol (iLTX), may result from the activity of epoxidehydrolase on LA [9
]. In normal fatty acid metabolism, LA is converted to GLA by means of a Δ6-desaturase in the rate-limiting step of essential fatty acid metabolism. GLA is then converted to DGLA and then to arachidonic acid (AA). In a single-dose study of radio-labeled LA metabolism in healthy subjects, the percent of LA recovered as AA was very low (0.05% of the radio-labeled dose) [10
In our literature review of central nervous system (CNS) effects related to GLA or LA, we found only two reports describing a total of five cases of purported GLA-induced seizures. Three patients with schizophrenia developed EEG-confirmed temporal lobe epilepsy after starting treatment with evening primrose oil. All patients improved after withdrawal of primrose oil and institution of carbamazepine treatment [11
]. Generalized tonic–clonic seizures were described in two further schizophrenic patients after starting treatment with evening primrose oil [12
]. However, one recent study has called these two earlier case series into question [13
]. It should be noted that all five cases used evening primrose oil rather than Borage oil, which was used by our patient. Borage oil may have the potential to lower the seizure threshold although some studies report anti-epileptic effects of fatty acids GLA and LA [4
]. Research in this regard is mixed and so no certain conclusions can be drawn. These herbal preparations may best be avoided by epilepsy patients at least until this issue is clarified [1
In a review of the toxicity of LA, the two monoepoxide metabolites (LTX diol and iLTXdiol) resulting from the action of epoxidehydrolase rather than Δ6-saturase, showed significant cytotoxic effects in a number of in vitro and in vivo test systems. Cytotoxic mechanisms included: reduced mitochondrial oxygen consumption and loss of cellular membrane potential. Organ effects included smooth muscle relaxation (porcine stomach), ARDS-like pulmonary injury (rats), and depressed cardiac function (dogs) [9
]. In addition, polyunsaturated fatty acids can be oxidized in cells and tissues to oxygenated α,β-unsaturated aldehydes [15
]. These putative toxins have been implicated in a number of pathological conditions including chronic inflammation, neurodegeneration, and cancer [16
]. Their role in seizures is unclear.
Analysis of a blood sample obtained on admission from our patient demonstrated an elevated concentration of GLA (1,165 μmol/L) but a low level of LA (871 μmol/L). Reference ranges (>18 years of age) for essential fatty acid analyses give a range of 16–150 μmol/L for GLA and 2,270–3,850 μmol/L for linoleic acid. (Information provided courtesy of Mayo Medical Laboratories, Rochester, MN, USA).
In a study of essential fatty acid levels in normal controls (n
31), asthma/allergic rhinitis patients (n
35), and atopic dermatitis patients (n
35), the control levels of LA were 2,902.65
742.25 μmol/L and control levels of GLA were 22.07
15.65 μmol/L [17
]. Another study of the pharmacokinetics of gamma-linolenic acid in healthy volunteers after evening primrose oil ingestion found levels of LA and GLA to be 327.3
155.2 μg/ml (1167.0
553.4 μmol/L) and 11.4
9.7 μg/ml (40.9
34.8 μmol/L), respectively [18
]. By contrast, in our patient, the non-lipid adjusted concentrations of LA and GLA were 244 μg/ml (871 μmol/L) and 324 μg/ml (1,163 μmol/L).
The reasons for the non-detection in CSF in our patient may include methodological limitations, reduction in GLA and LA concentrations due to a delay in lumbar puncture (which occurred 2 days after the blood sample that was analyzed for GLA and LA) and the possibility that an unmeasured toxic metabolite of GLA or LA (such as an oxygenated α,β-unsaturated aldehyde or diol) was present in CSF in greater quantity compared to GLA or LA. GLA and LA, along with a host of other polyunsaturated free fatty acids (palmitic, oleic, myristic) are present in normal CSF: a study of traumatic brain injury and healthy controls found a mean concentrations of LA in spinal fluid of 356 μmol/L (error range not provided) in controls [19
]. The presence of detectable levels of polyunsaturated free fatty acids in CSF in healthy individuals suggests methodologic failure as the likeliest explanation for the undetectable concentrations of GLA and LA in that matrix in our patient.
The ingestion of borage seed oil likely explains the high blood level of GLA detected in our patient. In normal fatty acid metabolism, dietary LA is converted to GLA by means of Δ6-saturase. Exogenous GLA, from ingestion of borage seed oil for example, would bypass this rate-limiting step and could lead to the very high level of GLA (relative to LA) found in our patient. The LA concentration in our patient was found to be lower than the LA concentrations found in control subjects during an atopic dermatitis–GLA treatment trial and in a pharmacokinetic study of GLA metabolism [17
]. Other explanations for the high levels of GLA and low levels of LA found in our patient could be unique to the product she ingested. Unfortunately, the herbal preparation which our patient started was never recovered and thus an analysis of the GLA and LA content could not be performed. A compounding or other formulation error could have resulted in excessive doses of GLA being administered. The high level of GLA found in our patient, the proximity of her consumption of borage seed oil 1 week before her seizures developed, and the limited case series describing temporal lobe seizures developing following GLA administration reported in the literature, suggest that consumption of borage seed oil was associated with our patient’s illness either by directly causing or unmasking an undiagnosed seizure disorder. It is equally possible that some as yet un-identified toxin in B. officinalis
could be responsible for the development of seizures either after excessive borage oil ingestion, or perhaps in a subset of those who use it therapeutically.
We attempted to obtain a complete list of supplements taken by the patient from her husband. Upon awakening from the pentobarbital-induced coma, it was difficult for the patient to provide any additional information because of short-term memory impairment. Two of her supplements, Coenzyme Q10 and L-carnitine have rarely been associated with seizures but not of the temporal lobe variety. It is possible that some non-reported supplement could have contributed to her illness either directly or via interacting with the borage oil.
Extensive clinical investigations in our patient failed to identify an alternate infectious or toxicological cause of her seizures. If anything, our case serves to highlight the uncertainties that surround herbal and CAM usage and the clinical and analytical challenges that face the medical toxicologist in approaching a possible case of botanical or herbal poisoning.