Letters to the Editor / Journal of Hepatology 49 (2008) 288–294 291

further light on the potential mechanism of hepatotoxic- References

ity in these sisters. Although it cannot be established for
certain if autoimmunity or a pharmaco-genetic prepon- [1] Manso G, López-Rivas L, Duque JM, Salgueiro E. Spanish reports
derance is the driving force behind these events, the reac- of hepatotoxicity associated with herbalifeÒ products. J Hepatol
2008;48:290–291.
tive hepatitis in the two sisters suggests a common [2] Elinav E, Pinsker G, Safadi R, Pappo O, Bromberg M, Anis E,
mechanism. Dr. Manso et al. raise the possibility that et al. Association between consumption of HerbalifeÒ nutri-
green tea or aloe vera ingredients, previously reported tional supplements and acute hepatotoxicity. J Hepatol
to induce hepatotoxicity may also be the cause in the 2007;47:514–520.
Spanish group of patients. However, these ingredients [3] Schoepfer AM, Engel A, Fattinger K, Marbet UA, Criblez D,
Reichen J, et al. Herbal does not mean innocuous: ten cases of
were not taken uniformly by all the reported patients severe hepatotoxicity associated with dietary supplements from
from Israel and Switzerland, and therefore a detailed list HerbalifeÒ products. J Hepatol 2007;47:521–526.
of all the ingredients including the chemical composition [4] Stickel F. Slimming at all costs: HerbalifeÒ-induced liver injury. J
is required as stated in the two original reports. In Spain Hepatol 2007;47:444–446.
as well as in Israel and Switzerland, it was not possible
to establish the true incidence of hepatotoxicity among
HerbalifeTM users since the number of consumers and Daniel Shouval
their demographic characteristics are only partially Eran Elinav
known. A rough calculation suggested an incidence of Hadassah-Hebrew University Medical Center,
25–30 cases per 100,000 HerbalifeTM products among Is- P.O. Box 12000, Jerusalem 91120, Israel
raeli consumers. We agree with the colleagues from Tel.: +972 2 6777 337; fax: +972 2 6420 338
Spain that the true incidence may be higher since liver (D. Shouval)
injury may remain occult and asymptomatic as is well E-mail address: shouval@cc.huji.ac.il (D. Shouval)
known by practising hepatologists. Thus, it is important
to draw the attention of the Journals readers to reports
of similar cases. doi:10.1016/j.jhep.2008.05.006

Herbalife nutritional products and liver injury revisited

To the Editor: reports with similar characteristics were documented

Recently, two reports of 12 and 10 cases of idio- over seven years.
pathic liver disease from Israel [1] and Switzerland Taken at face value, these few cases represent an ex-
[2], respectively, appeared in the same issue of this tremely low incidence of suspected liver injury among
journal alleging in the articles and in an accompanying the millions of Herbalife consumers worldwide (5.5 mil-
editorial that these cases were caused by the consump- lion consumers in 2004 alone). In 2004 more than
tion of Herbalife products. In fact, the two papers [1,2] 40,000,000 servings from 29 different products were dis-
concluded unequivocally that causation was certain in tributed to approximately 37,000 Israeli consumers. A
some cases. Despite the authors’ strong assertion that similar number of servings from 26 different products
a cause and effect relationship between Herbalife prod- were distributed to approximately 80,000 Swiss consum-
uct use and liver injury exists, an objective review of ers. In the 5 combined cases where causation was
the facts raises serious doubt whether such a conclu- assessed as ‘‘certain” due to a reported combination of
sion can be drawn. In Israel [1], the observed liver positive dechallenge and rechallenge, the case details
abnormalities resolved in eleven of the twelve cases re- remain unclear. As an example, 4 of these patients
ported. The remaining patient succumbed to complica- described between 3 and 17 different and specific Herba-
tions of liver transplantation and despite the fact that life products previously consumed prior to dechallenge,
this patient also had evidence of hepatitis B infection, and in the remaining patient no specific Herbalife prod-
sole attribution of the liver injury to Herbalife products uct in either the reported dechallenge or rechallenge
was made. Another patient had stage I primary biliary could be identified. Equally concerning is the fact that
cirrhosis and other patients consumed a variety of sub- there is no mention of which of the initially reported
stances or had other co-morbid diseases that could products were subsequently consumed by any of the pa-
have caused or contributed to the liver disease ob- tients that allegedly led to the recurrence of symptoms.
served. In Switzerland [2], a total of 10 evaluable case Even if all 4 patients consumed the exact same 3–17
292 Letters to the Editor / Journal of Hepatology 49 (2008) 288–294

products, respectively, there was no purported toxic since 1992 when this survey was done, the number of
ingredient identified by the authors in the article that idiopathic liver diseases remains a significant percent-
was common among all 4 of these patients. It is also age of all the cases.
interesting that in one of these 4 patients, rechallenge Herbalife nutritional products are registered and
of unspecified Herbalife product was reported to result notified as foods, meal replacements, and dietary sup-
in recurrence of symptoms, yet the patient purportedly plements and not as herbal medicines. There are ingre-
continues to use unspecified Herbalife product(s) and re- dients such as guarana, green tea, and caffeine, which
mains asymptomatic. are being used extensively in numerous food products
The generally accepted criteria for causality of liver and are not unique to Herbalife. Herbalife conducts
diseases cited in the editorial, including dechallenge/ testing, through independent laboratories, on product
rechallenge, were designed to consider drug-induced batches for heavy metals, pesticides, ochratoxin A,
hepatotoxicity when a specific defined ingredient has aflatoxins, comfrey retrosine (pyrrolizidine alkaloids;
been identified, and were not designed to consider PAs), and kava kavalactones. Completed tests have
the effects of multiple different foods, supplements, consistently shown no detection of pesticides, kava or
and distinct nutritional products in combination or PAs, and traces of ochratoxins, aflatoxins, and heavy
totality. The comment made in the editorial [3] that metals are below minimum threshold levels. Also, the
‘‘there can be little doubt that these products were company has a well-developed adverse event reporting
the cause” is simply not supported by the facts. In system which monitors and evaluates adverse events
acute hepatotoxicity, liver injury typically occurs with globally and has sought the involvement of outside,
a substantial and predictable frequency, its severity is independent experts to evaluate its adverse event
dose-dependent, and a responsible agent can be iden- experience.
tified. In contrast, liver injury from an immune-medi- The core products of Herbalife provide healthy solu-
ated hypersensitivity reaction is sporadic, and clinical tions to the worldwide epidemic of obesity. The prod-
symptoms and abnormal liver biochemical tests occur ucts are designed to deliver balanced nutrition and
in only a very small number of individuals who met- assist in the promotion of fitness and a healthy, active
abolically convert some constituent of the product lifestyle. These products are primarily based in vegeta-
consumed into a substance that stimulates an immune ble proteins, fish oils, vitamins, and minerals for which
reaction [4]. While this is the generally accepted sce- safe use is very well established. Some of the com-
nario among hepatologists, the exact cause, predispos- pany’s products also contain botanical ingredients that
ing individualistic factors, and precise pathophysiology are well characterized and tested. These botanicals are
of this rare form of liver disease remain poorly included in Herbalife product at levels that are in a
understood. safe nutritional range where they have antioxidant
Furthermore, the incidence of these cases appears to properties and support normal function, and are la-
be very low compared to the risks associated with some beled in accordance with all the applicable laws. The
over-the-counter and prescription drugs on the market. company does not market or encourage the use of
A population-based survey in Atlanta, Georgia, USA any of its products as medicines for the treatment of
published in 2007 found that the incidence of acute li- specific diseases.
ver failure in eight counties was 5.5 per million individ- There were no undefined or unlabeled herbs in these
uals, and the use or abuse of acetaminophen was products as suggested in the articles and contrary to
associated with 41 percent of the cases, while a signif- what has been portrayed, the company has cooperated
icant percentage of adults had liver failure of unknown fully with the ministries of health in their investigation
etiology [5]. As indicated in the Swiss article, the rare of these cases. In fact, government officials and clini-
incidence of these events having been ten cases col- cians investigating these cases in both countries were gi-
lected over seven years, makes the overall incidence ven access to full product formulas and ingredients.
some 1.8 per million patients/year [2]. This low level These ingredient disclosures were documented in numer-
of risk of liver disease is indistinguishable from the ous communiqués and included full product dossiers
background incidence of idiopathic liver disease, where and results of independent testing regarding product
no cause can be found. In a study of 71,000 North purity and integrity.
Americans in 1992, the background rate of idiopathic It is also unfortunate in our view that Herbalife’s
or cryptogenic liver disease was 24 cases per 100,000 brand name was generically linked to liver injury
individuals compared to 14 per 100,000 attributed to rather than specific products or ingredients where
cases of hepatitis B, 25 per 100,000 due to alcoholism, such an association could not be established. This ap-
and 7 per 100,000 due to other viral illnesses [6]. While proach is unprecedented. Although Herbalife remains
the spectrum of liver diseases has certainly changed committed in the spirit of product stewardship to
 Letters to the Editor / Journal of Hepatology 49 (2008) 288–294 293

cooperate openly and fully in the evaluation of any

Louis Ignarro
adverse event potentially associated with a Herbalife
David Heber à
product, we believe that objective review of these
David Geffen School of Medicine,
cases does not support a cause and effect relationship
UCLA, 900 Veteran Avenue,
between any specific Herbalife product or ingredient
Room 12217-Warren Hall,
and liver injury.
Los Angeles, CA 90095, USA

Y. Steven Henig
References
Ezra Bejar
[1] Elinav E, Pinsker G, Safadi R, Pappo O, Bromberg M, Anis E,
Herbalife International,
et al. Association between consumption of Herbalife nutritional 990 West 190th St. Torrance,
supplements and acute hepatotoxicity. J Hepatol CA 90502, USA
2007;47:514–520. Tel.: +1 310 410 9600; fax: +1 310 767 3314 (E. Bejar)
[2] Schoepfer AM, Engel A, Fattinger K, Marbet UA, Criblez D, E-mail address: ezrab@herbalife.com (E. Bejar)
Reichen J, et al. Herbal does not mean innocuous: 10 cases of
severe hepatotoxicity associated with dietary supplements from
Herbalife products. J Hepatol 2007;47:521–526.
[3] Stickel F. Slimming at all costs: Herbalife-induced liver injury. J Dr. Ignarro is a member of the Scientific Advisory Board of Herbalife
Hepatol 2007;47:444–446. International. Dr. Ignarro’s title is for identification purposes only. The
[4] Gunawan B, Kaplowitz N. Clinical perspectives on xenobiotic- University of California does not endorse specific products or services as
induced hepatotoxicity. Drug Metab Rev 2004;36:301–312. a matter of policy.
à
[5] Bower WA, Johns M, Margolis HS, Williams IT, Bell BP. Dr. Heber is Chairman of the Nutritional and Scientific Advisory
Population-based surveillance for acute liver failure. Am J Boards of Herbalife International. Dr. Heber’s title is for identification
Gastroenterol 2007;102:2459–2463. purposes only. The University of California does not endorse specific
[6] Walker AM, Cavanaugh RJ. The occurrence of new hepatic products or services as a matter of policy.
disorders in a defined population. Post Marketing Surveillance
1992;6:107–117. doi:10.1016/j.jhep.2008.05.005

Herbalife revisited: Reply

To the Editor: HerbalifeTM associated hepatotoxicity could recently be

We read with interest the HerbalifeTM response by Dr. made, following information requested from Herbalife
Ignarro and co-authors to the two reports and Editorial by the Israeli ministry of health and received in 2007.
printed in the Journal on association of HerbalifeTM with An estimated incidence of 25–30 cases per 100,000 con-
hepatotoxicity. We wish to address a number of state- sumers was made. This figure is only an approximation,
ments in their letter: since demographic data on the population of consumers
was unavailable at time of analysis. We also disagree with
1. Incidence the authors of the letter that the so called ‘‘low level of risk
of liver disease is indistinguishable from the background
We agree with the authors that the incidence of Herba- incidence of idiopathic liver disease”. In our survey in
lifeTM-associated hepatotoxicity is probably low, but not all Israeli hospitals, we identified initially 12/33 hospital-
as low as they suggest. We disagree with their comment ized patients with liver injury of so called undetermined
that 22 cases among 5.5 million consumers world-wide etiology who reported intake of HerbalifeTM products.
can be used as proof for a low incidence of the com- This still leaves 21/33 patients with so called idiopathic li-
pound(s) associated hepatotoxicity. The cases reported ver disease reflecting an incidence of <6 cases/million of
by us, were identified through an ICD-9 search in all Israe- undetermined etiology of the liver injury (after exclusion
li hospitals during a two year period, starting in 2004. This of HerbalifeTM consumers). Finally it is common knowl-
survey identified 12 cases which reported intake of Herba- edge among hepatologists that sub-clinical, asymptom-
lifeTM products among 33 patients diagnosed with crypto- atic ALT elevation may occur in patients with occult
genic liver dysfunction. The association between intake of liver disease. Our survey included only hospitalized pa-
HerbalifeTM products and hepatic injury was classified as tients, identified retrospectively through hospital records.
certain in 3, probable in 6 and possible in 3 patients using Therefore, the number of patients who may have devel-
WHO criteria. A rough calculation of the incidence of oped occult hepatotoxicity in association with HerbalifeTM