ARTICLE IN PRESS

Phytomedicine 13 (2006) 487–493

www.elsevier.de/phymed

Simultaneous analysis of artemisinin and flavonoids of several extracts

of Artemisia annua L. obtained from a commercial sample and
a selected cultivar
A.R. Biliaa,, P. Melillo de Malgalhaesb, M.C. Bergonzia, F.F. Vincieria
a
Department of Pharmaceutical Sciences, University of Florence, Via U. Schiff 6, 50019 Sesto Fiorentino-Florence, Italy
b
Divisão de Agrotecnologia, CPQBA-UNICAMP, C.P. 6171, 13.081.970 Campinas, SP, Brasil

Abstract

Artemisia annua L. (Qinghao) is a promising and potent antimalarial herbal drug. This activity has been ascribed to
its component artemisinin, a sesquiterpene lactone that is very effective against drug-resistant Plasmodium species with
a low toxicity. Our studies indicate that several flavonoids of A. annua can promote and enhance the reaction of
artemisinin with hemin. These data are in good agreement with previous investigations on the in vitro potentiation of
antimalarial activity of artemisinin by such flavonoids.
As a consequence, in view of a possible use of the phytocomplex rather than pure artemisinin, an HPLC/DAD/MS
method is proposed for the simultaneous detection and quantification of both flavonoids and artemisinin. Different
extracts, obtained from two different herbal drugs, a commercial sample and a selected cultivar, were analyzed in order
to determine which solvents provide the best yields of both artemisinin and flavonoids. Qualitative and quantitative
results obtained using an HPLC method are described, which will be useful for developing highly effective herbal drug
preparations.
r 2006 Elsevier GmbH. All rights reserved.

Keywords: Artemisia annua L.; Commercial sample and selected cultivars; Extracts; Artemisinin; Flavonoids; HPLC/DAD/MS

Introduction sesquiterpenoids, flavonoids, coumarins, triterpenoids,

Artemisia annua L. (sweet or annual wormwood,
family Asteraceae) is an annual herb endemic to the
northern parts of Chahar and Suiyuan provinces in
China where it is known as ‘quinghao’ (green herb) and
it has been used to treat chills and fever for more than
2000 years (Klayman, 1985; Hien and White, 1993).
A search for the active compounds responsible for the
activity of A. annua has led to the isolation of
Corresponding author. Tel.: +39 0554573708;
fax: +39 055457379.
E-mail address: ar.bilia@unifi.it (A.R. Bilia).
steroids, phenolics, purines, lipids and aliphatic com-
pounds from different plant parts (Bhakuni et al., 2001).
Among these metabolites, artemisinin (quinghaosu) is
considered the main active constituent (Klayman, 1985;
van Agtmael et al., 1999; Heppner and Ballou, 1998). Its
level can vary considerably, depending on the plant
material, growing conditions and seasonal and geo-
graphic variations: it is generally present in the leaves
and flowers of the plant at 0.01–1.4% dry weight
(Delabays, 1997). Artemisinin is considered a potent
antimalarial drug, even against chloroquine- and qui-
nine-resistant Plasmodium falciparum and other malaria-
causing parasites. Its activity is based on an unusual

0944-7113/$ - see front matter r 2006 Elsevier GmbH. All rights reserved.
doi:10.1016/j.phymed.2006.01.008
 ARTICLE IN PRESS
488 A.R. Bilia et al. / Phytomedicine 13 (2006) 487–493

mechanism of action, still the object of much debate, R3

which most likely involves the formation of free-radical
intermediates, originating from the direct interaction of R2
the endoperoxide group with the heme iron and leading
to the alkylation of malarial-specific proteins (Heppner
and Ballou, 1998). H3CO O
Malaria continues to be a major health problem in
tropical and subtropical areas (World Health Organiza-
tion and The World Health Report, 1996) and the
H3CO R1
emergence of chloroquine-resistant strains has led to the
search for new antimalarial drugs. Artemisinin is a
OH O
promising candidate. It offers impressive effects, includ-
ing ‘high efficacy, fast-action and low-toxicity’, and is
regarded as ‘a breakthrough in the history of antimalar- R1 R2 R3
ial drugs’ (Heppner and Ballou, 1998). It is a potent
1. Chrysosplenol-D -OCH3 -OH -OH
blood schizonticide with a minimum inhibitory concen-
tration of 10
7 M (De Vries and Dien, 1996). 2. Eupatin -OH -OCH3 -OH
Artemisinin is now available commercially in China, 3. Cirsilineol -H -OH -OCH3
Vietnam and other countries as an antimalarial drug
4. Casticin -OCH3 -OCH3 -OH
and, although its total synthesis has been achieved (the
novo synthesis gives low yields and is uneconomical due 5. Chrysoplenetin -OCH3 -OH -OCH3
to its complex structure), isolation from the plant still 6. Artemetin -OCH3 -OCH3 -OCH3
represents the best alternative (Klayman, 1985; De Vries
and Dien, 1996). Fig. 1. Chemical structure of flavonoids of the Artemisia
Over the last two decades it has also been suggested annua L. extract.
that the efficacy of the A. annua plant itself derives from
a synergistic effect and that it is a combination of In the present paper, we report the HPLC/DAD/MS
constituents in the plant which produce the total anti- analyses of different extracts of the aerial parts of two
plasmodial activity. Several polymethoxyflavones such samples of A. annua L.: a Chinese commercial sample
as casticin, artemetin, chrysosplenetin, chrysosplenol-D (named AA) and a cultivar with high artemisinin
and circilineol (Bhakuni et al., 2001) may contribute to content, selected by the University of Campinas (named
the in vitro activity of artemisininin against P. AAB). Artemisia annua L. is an example of a phytome-
falciparum (Bhakuni et al., 2001; Elford et al., 1987; dicine for which a whole or partially purified extract
Phillipson, 1999; Liu et al., 1992). However, data offers advantages over single isolated constituents
concerning the mechanism related to the synergistic (Williamson, 2001), and we have thus attempted to
effect have not yet been reported. develop a chromatographic method to detect and
We have recently demonstrated, by simple and rapid quantify artemisinin and other constituents of the
physical–chemical methods, that the flavonoids play an phytocomplex simultaneously. This analytical method
important role in the interaction between artemimisin was used to evaluate the content of constituents in
and hemin (Bilia et al., 2002). The studies involved hexane and dichloromethane extracts obtained by
chrysoplenetin, casticin, eupatin and chrysosplenol-D maceration of the herbal drugs.
(Fig. 1), isolated by repeated column chromatography
from A. annua L. and the interaction between artemi-
sinin and hemin, evaluated by UV/Vis and HPLC/
DAD/MS analysis, demonstrated that the synergism of Material and methods
flavonoids is related to the assisted activation of
artemisinin before its interaction with hemin (Bilia et Chemicals
al., 2002).
The potential use of extracts of A. annua L. instead of Commercial aerial parts of A. annua L. (500 g) were
pure artemisinin, particularly in areas where large-scale kindly offered by Xu Dong Mei, Liaoning Huaxi Group
pharmaceutical production is not possible, is clearly of (Corporation Dalian, China). A sample (300 g) of a
interest, not least for economic reasons (Pras et al., selected high yield cultivar of A. annua L. was provided
1991). However, the concentration of artemisinin in the by P.M.M. of the Universidade Estadual de Campinas
plant is very low and the most successful way to increase (Brazil). The Brazilian hybrid plant was obtained
artemisinin content is through the breeding of high- according to the procedure carried out by MEDI-
yielding plants. PLANT (Magalhães et al., 1997).
 ARTICLE IN PRESS
A.R. Bilia et al. / Phytomedicine 13 (2006) 487–493
Artemisinin was purchased from Sigma (Sigma-
Aldrich S.r.l., Milan, Italy)
Chrysosplenol-D and eupatin were isolated as pure
constituents from the commercial herbal drug, along
with chrysoplenetin and casticin in an inseparable
mixture. Both the pure constituents and the mixture
were identified by NMR, as reported previously (Bilia et
al., 2002). Chrysosplenol-D was used for the calibration
of the polymethoxylated flavonoids.
All the solvents used for the extraction and HPLC
analysis (MeOH, hexane, dichloromethane, and acet-
onitrile) were HPLC grade from Merck (Darmstadt,
Germany); 85% formic acid was provided by Carlo
Erba (Milan, Italy). Water was purified using a Milli-
Qplus system (Milford, MA, USA).
Preparation of the extracts
The dried aerial parts of both sweet wormwood
samples were cut into small pieces with an Osterizer.
Samples of 10 g of material were extracted exhaustively
at room temperature by maceration with 100 ml of
dichloromethane or hexane for 72 h. The eluates were
subsequently dried in vacuo to obtain the crude extracts.
HPLC-DAD and HPLC-MS systems
The HPLC analyses were performed using a HP 1100
Liquid Chromatograph (Agilent Technologies, Palo
Alto, CA, USA) equipped with a HP 1040 diode-array
detector (DAD), an automatic injector, an auto sampler
and a column oven and managed by a HP 9000
workstation (Agilent Technologies, Palo Alto, CA,
USA).
The HPLC system was interfaced with a HP 1100
MSD API-electrospray (Agilent Technologies, Palo
Alto, CA, USA). The interface geometry, with an
orthogonal position of the nebulizer with respect to
the capillary inlet, allowed the use of analytical
conditions similar to those of HPLC-DAD analysis.
Mass spectrometry operating conditions were optimized
in order to achieve maximum sensitivity values: gas
temperature 350 1C at a flow rate of 10 l/min, nebulizer
pressure 30 psi, quadrupole temperature 30 1C, and
capillary voltage 3500 V. Full scan spectra from m/z
100 to 800 in the positive-ion mode were obtained (scan
time 1 s).
Separations were performed on a reversed-phase
column Purosphers Star RP-18, namely Hibars Pre-
packed column RT (250  4.6 mm) with a particle size of
5 mm (Merck, Darmstadt, Germany). The eluents were:
water adjusted to pH 3.2 by formic acid (A), and
acetonitrile (B). The mobile phase was isocratic, using
50% A and 50% B for 20 min at a flow rate of 1.3 ml/
489
min. The system was operated with oven temperature at
26 1C; the injection volume was 20 ml.
Before HPLC analysis, each sample was filtered
through a cartridge-type sample filtration unit with a
polytetrafluoroethylene (PTFE) membrane (d ¼ 13 mm,
porosity 0.45 mm, Lida manufacturing Corp.) and
injected immediately.
Chromatograms were recorded at 280 nm for the
flavonoids and at 210 nm to detect artemisinin and any
other constituents. DAD spectra between 200 and
450 nm were stored for all peaks exceeding a threshold
of 0.1 mAu.
HPLC-DAD method
Calibration and method validation
Calibration curves were obtained from a methanolic
dilution series of the reference standard containing both
artemisinin and chrysosplenol-D; 20.0 mg of both
compounds were placed in a 10-ml volumetric flask
and the volume was brought to 10 ml with methanol.
Concentrations ranging between 0.01 and 2.0 mg/ml
of chrysosplenol-D and 0.1 and 2.0 mg/ml of artemisinin
were prepared from the standard stock solution by serial
dilution with methanol for the calibration data. All
calibration levels (n ¼ 5) were measured six times.
Linear regression was used to establish the calibration
curve. Results were calculated using the peak areas.
Three samples of 10 mg of dried n-hexane extract
(AAB sample accurately weighed) were suspended in
2 ml acetonitrile in a volumetric flask and sonicated for
20 min. The suspensions were filtered and each of the
sample solutions was injected three times. The artemi-
sinin and flavonoid content was calculated on the basis
of the peak area.
Sample analysis
Five milligrams of the dichloromethane and hexane
dried extracts were accurately weighed and suspended in
acetonitrile (1.0 ml) in a volumetric flask. The suspen-
sions were sonicated for 20 min and filtered through a
cartridge-type sample filtration unit before HPLC
analysis.
Results and discussion
Preparation of the extracts
Liquid solvent extraction with toluene (Laughlin,
1994), n-hexane (Dharam et al., 1996; Elsohly et al.,
1987), chloroform (Tang et al., 2000; Woerdenbag et al.,
 ARTICLE IN PRESS
490 A.R. Bilia et al. / Phytomedicine 13 (2006) 487–493

1991), petroleum ether (Klayman et al., 1984), or The dichloromethane extract from AAB was em-
dichloromethane (Brown et al., 2003; Sy and Brown, ployed in order to develop the analytical method.
2001) is currently the most common technique used for
artemisinin extraction. Because n-hexane, toluene and
petroleum ether showed no significant differences when HPLC-DAD and HPLC-MS analysis of the
used for the extraction of artemisinin (Vandenberghe et extracts
al., 1995), dichloromethane and hexane were tested in
our investigation in order to determine the best solution A number of analytical methods exist for the
for recovery of both artemisinin and methoxylated evaluation of artemisinin in plant extracts, but none
flavonoids. Preparation of these extracts is reported in has been proposed specifically for the analysis of
the experimental section.Two different samples of the the methylated flavonoids. The HPLC/DAD/MS
herbal drug, namely AA (commercial Chinese A. annua) method was determined by varying the flow rate
and AAB (A. annua from Brazil), were used. (1.0, 0.8 and 0.6 ml/min) and using either a binary

MSD TIC, API-ES, Pos, Scan, 120

20000000
2.23

17500000

4.909
15000000 2.917 15.274
8.124
9.213
12500000

6.657
10000000
5.989 7.42 10.580
3
7500000
14.051
17.929
11.499
5000000

2500000

2 4 6 8 10 12 14 16 18
min

time=15.144 API-ES, Pos, Scan, 120

100 305.4
Max: 1.45408e+006

80

60

151.2

40

20
306.4
13
7. 587.9
1
0

200 400 600 800

m/z

Fig. 2. Total ion current (TIC) of CH2Cl2 extract of Artemisia annua L. extract and mass fragmentation of the peak of artemisinin.
 ARTICLE IN PRESS
A.R. Bilia et al. / Phytomedicine 13 (2006) 487–493 491

(water and methanol) or ternary (water, methanol and
acetonitrile) gradient. No significant differences were
found when the flow was varied, while binary and
ternary eluents greatly affected the tailing factor of the
separation. Using the mobile phase with water and
acetonitrile provided the best peak shape and the best
separation of artemisinin from the flavonoids in the
extracts.
The developed analytical system made it possible
to separate artemisinin (1, Rt ¼ 15.1) and flavo-
noids clearly. Detection of constituents was performed
both at 210 and 280 nm and artemisinin was identified as
the peak at 15.1 min using mass spectrometry, a very
efficient method for detection, particularly in cases of
constituents without chromophore moieties. Thus, the
registered total ion current (TIC) of the dichloro-
methane extract and the mass spectrum of artemisinin
are reported in Fig. 2, where the quasi-molecular
fragment, i.e., [M+Na]+ ¼ 305 m/z, of the peak
corresponding to artemisinin is shown.
Identification of flavonoid peaks was obtained
by combining the HPLC/DAD and HPLC/MS
and comparison with previous reports (Bilia et al.,
2002). UV spectra of each peak and the charac-
teristic fragments in the MS spectrum were useful for
identifying all the flavonoids. Peaks of eupatin,
chrysosplenol-D and the unseparable mixture of casticin
and chrysoplenetin were also identified by the compar-
ison of their retention times with those of the isolated
compounds. The presence of cirsilenol and artemetin
was confirmed using literature data (Bhakuni et al.,
2001). The chromatogram registered at 280 nm; Fig. 3
shows the flavonoid peaks for the dichloromethane
extract.

4.87
1. Chrysosplenol-D, [M + Na]+ = 383 m/z
700
2 2. Eupatin, [M + H]+ = 361 m/z
2.88 4, 5 3. Cirsilineol, [M + H]+ = 345 m/z
600 1 8.08
4. Casticin, [M + H]+ = 375 m/z
5. Chrysoplenetin, [M + H]+ = 375 m/z
500 6. Artemetin, [M + H]+ = 389 m/z

4.60
400
mAU

300

200

6
100
14.00
3
0

0 2 4 6 8 10 12 14 16 18
min

Fig. 3. Chromatogram at 280 nm of the CH2Cl2 extract of Artemisia annua L.

Table 1. Extraction yields and percentage of artemisinin and flavonoids in the tested extracts

Extract samplesa Extraction yield (g)b Artemisinin content Flavonoid content

(%7SEM)c (%7SEM)c

Hexane AAB 3.86 3.01 (1.14) 2.82 (1.25)

Hexane AA 2.30 0.87 (0.17) 0.50 (0.54)
Dichloromethane AAB 3.58 1.04 (0.78) 5.72 (1.14)
Dichloromethane AA 1.97 0.12 (0.98) 0.68 (1.20)
a
AA: Chinese commercial sample of Artemisia annua; AAB: cultivar of Artemisia annua selected by the University of Campinas.
b
Dried extract obtained from 10 g herbal drug.
c
Percentages w/w are referred to the extract weight, n ¼ 3 samples.
 ARTICLE IN PRESS
492 A.R. Bilia et al. / Phytomedicine 13 (2006) 487–493
Quantitation of artemisinin and flavonoids
Quantitation of constituents was performed using UV
detection, at 210 nm for artemisinin and 280 nm for
flavonoids, using chrysosplenol-D as standard. The calibra-
tion range of artemisinin content wsa between 0.1 and
2 mg/ml. and that of chrysosplenol-D was between 0.01 and
2 mg/ml. Examination of system precision (reproducibility)
showed 0.45% relative standard deviation for artemisinin
and 0.65% relative standard deviation for chrysosplenol-D
on the basis of six injections of the same calibration
solution. The limit of quantitation (LoQ) for artemisinin
was 0.5 mg, the limit of detection (LoD) was 0.3 mg. The
calibration function of artemisinin showed a linear response
of the diode array detector (Y ¼ 171:84X þ 4:8 and
r2 ¼ 0:9999) as was also true for chrysosplenol-D
(Y ¼ 1639:6X þ 328:85 and r2 ¼ 0:9974).
Very different yields and contents of artemisinin and
flavonoids were found in the tested extracts, as reported in
Table 1. There was also a deep variability between the two
samples of sweet wormwood AA and AAB, with a
considerably increased flavonoid and artemisinin content
in the Brazilian cultivar, a breeding of high yielding plants
which contains a larger quantity of artemisinin. As
reported in Table 1, both n-hexane and dichloromethane
can selectively extract artemisinin, as well as flavonoid
constituents. Considering the percentage and total amount
of artemisinin in the extracts from AA and AAB samples,
the latter plant material represents a rich source of the
phytocomplex, both artemisinin and flavonoids.
Conclusion
Malaria is the leading cause of disease in many parts of
the world, and the effects of the disease are further
compounded by the extremely difficult living conditions in
some of these poorest nations. The rapid increase of
multi-drug resistant parasites has rendered previous
malarial control methods impractical, risky and no longer
cost effective. Artemisia annua L. could be an effective,
inexpensive and easily accessible drug in these regions.
The content of constituents of A. annua L., especially the
active principle artemisinin, is strongly affected by growth
conditions, seasonal and geographic variations and
breeding. Authorities in affected regions are conducting
research to find high-yielding cultivars and develop them
for in situ cultivation. It has been demonstrated that in
addition to artemisinin, flavonoids are also involved in the
plant’s antimalarial activity; thus, it is important that any
technique used for the analysis of plant samples and
extracts is able to detect both types of constituents in
order to assess quality and stability.
The analytical procedure we have developed in the
present work led to the quantification of both flavonoids
and artemisinin in extracts of A. annua L. by combining
HPLC-DAD and HPLC-MS. The latter method was
also very selective for the detection of artemisinin in
trace amounts.
The selected Brazilian cultivar contains a higher level
of phytocomplex with respect to the commercial Chinese
herbal drug. n-Hexane was more selective in isolating
artemisinin, while dichloromethane was more selective
for the extraction of flavonoids.
Acknowledgements
This work was supported in part by Ministero
dell’Istruzione, dell’Università e della Ricerca (MIUR,
Rome, PRIN 2004). The authors are grateful to Dr.
Sandra Gallori for her expert assistance with the
chromatographic aging experiments.
References
Bhakuni, R.S., Jain, D.C., Sharma, R.P., Kumar, S., 2001.
Secondary metabolites of Artemisia annua and their
biological activity. Curr. Sci. 80, 35–48.
Bilia, A.R., Lazari, D., Messori, L., Taglioli, V., Temperini,
C., Vincieri, F.F., 2002. Simple and rapid physico-chemical
methods to examine action of antimalarial drugs with
hemin: its application to Artemisia annua constituents. Life
Sci. 70, 769–778.
Brown, G.D., Liang, G.Y., Sy, L.K., 2003. Terpenoids from
the seeds of Artemisia annua. Phytochemistry 64, 303–323.
De Vries, P.J., Dien, T.K., 1996. Clinical pharmacology and
therapeutic potential of artemisinin and its derivates in the
treatment of malaria. Drugs 52, 818–836.
Delabays, N. Biologie de la reproduction chez l’Artemisia
annua L. et génetique de la production en Artemisinine.
Ph.D. Thesis, University of Lausanne, 1997.
Dharam, C.J., Ajay, K.M., Madan, M.G., Anil, K.S., Ram,
K.V., Ajai, P.G., Sushil, K., 1996. Isolation of high
artemisinin-yielding clones of Artemisia annua. Phytochem-
istry 43, 993–1001.
Elford, B.C., Roberts, M.F., Phillipson, J.D., Wilson, R.J.,
1987. Potentiation of the antimalarial activity of qinghaosu
by methoxylated flavones. Trans. R. Soc. Trop. Med. Hyg.
81, 434–436.
Elsohly, H.N., Croom, E.M., Elsohly, M.A., 1987. Analysis of
the antimalarial sesquiterpene artemisinin in Artemisia
annua by high-performance liquid chromatography
(HPLC) with postcolumn derivatization and ultraviolet
detection. Pharm. Res. 4, 258–260.
Heppner, D.G., Ballou, W.R., 1998. Malaria in 1998:
Advances in Diagnosis, Drugs and Vaccine Development,
Tropical and Travel Association.
Hien, T.T., White, N.J., 1993. Qinghaosu. Lancet 341,
603–608.
Klayman, D.L., 1985. Qinghaosu (artemisinin): an antimalar-
ial drug from China. Science 228, 1049–1055.
 ARTICLE IN PRESS
A.R. Bilia et al. / Phytomedicine 13 (2006) 487–493
Klayman, D.L., Lin, A.J., Acton, N., Scovill, J.P., Hoch, J.M.,
Milhous, W.K., Theoharides, A.D., Dobek, A.S., 1984.
Isolation of artemisinin (qinghaosu) from Artemisia annua
growing in the United States. J. Nat. Prod. 47, 715–717.
Laughlin, J, 1994. Agricultural production of artemisinin – a
review. Trans. R. Soc. Trop. Med. Hyg. 88, 21.
Liu, K.C.S.C., Yang, S.L., Roberts, M.F., Elford, B.C.,
Phillipson, J.D., 1992. Antimalarial activity of Artemisia
annua flavonoids from whole plants and cell cultures. Plant
Cell Rep. 11, 637.
Magalhães, P.M., Delabays, N., Sartoratto, A., 1997. Ciencia
e Cultura. J. Braz. Assoc. Adv. Sci. 49(5/6).
Phillipson, J.D., 1999. New drugs from nature – it could be
yew. Phytother. Res. 13, 2.
Pras, N., Visser, J.F., Batterman, S., Woerdenbag, H.J., Maligré,
T.M., Lugt, C.B., 1991. Laboratory selection of Artemisia
annua L. for high yielding types. Phytochem. Anal. 2, 80–83.
Sy, L.K., Brown, G.D., 2001. Deoxyarteannuin B, dihydro-
deoxyarteannuin B and trans-5-hydroxy-2-isopropenyl-5-
methylhex-3-en-1-ol from Artemisia anuua. Phytochemistry
58, 1159–1166.
493
Tang, HQ., Hu, J., Yang, L., Tan, R.X., 2000. Terpenoids and
flavonoids from Artemisia species. Planta Med. 66,
391–393.
van Agtmael, M.A., Eggelte, T.A., van Boxtel, C.J.,
1999. Artemisinin drugs in the treatment of malaria:
from medicinal harb to registered medication. TiPS 20,
199–205.
Vandenberghe, D.R., Vergauwe, A.N., Van Montagu, M.,
Van Den Eeckhout, E.G., 1995. Simultaneous determina-
tion of artemisinin and its bioprecursors in Artemisia annua.
J. Nat. Prod. 58, 798–803.
Williamson, E.M., 2001. Synergy and other interactions in
phytomedicines. Phytomedicine 8 (5), 401.
Woerdenbag, H.J., Pras, N., Bos, R., Visser, J.F., Hendricks,
H., Malingré, T.M., 1991. Analysis of artemisinin and
related sesquiterpenoids from Artemisia annua L. by
combined gas chromatography/mass spectrometry. Phyto-
chem. Anal. 2, 215–219.
World Health Organization, The World Health Report, 1996.
Fighting disease, fostering development. WHO, Geneva,
Switzerland.