An ethnobotanical survey of medicinal plants used in the East Sepik province of Papua New Guinea

Background Rapid modernization in the East Sepik (ES) Province of Papua New Guinea (PNG) is resulting in a decrease in individuals knowledgeable in medicinal plant use. Here we report a synthesis and comparison of traditional medicinal plant use from four ethnically distinct locations in the ES Province and furthermore compare them to two other previous reports of traditional plant use from different provinces of PNG. Methods This manuscript is based on an annotated combination of four Traditional Medicines (TM) survey reports generated by University of Papua New Guinea (UPNG) trainees. The surveys utilized a questionnaire titled “Information sheet on traditional herbal preparations and medicinal plants of PNG”, administered in the context of the TM survey project which is supported by WHO, US NIH and PNG governmental health care initiatives and funding. Regional and transregional comparison of medicinal plant utilization was facilitated by using existing plant databases: the UPNG TM Database and the PNG Plant Database (PNG Plants) using Bayesian statistical analysis. Results Medicinal plant use between four distinct dialect study areas in the ES Province of PNG showed that only a small fraction of plants had shared use in each area, however usually utilizing different plant parts, being prepared differently and to treat different medical conditions. Several instances of previously unreported medicinal plants could be located. Medicinally under- and over-utilized plants were found both in the regional reports and in a transregional analysis, thus showing that these medicinal utilization frequencies differ between provinces. Conclusions Documentation of consistent plant use argues for efficacy and is particularly important since established and effective herbal medicinal interventions are sorely needed in the rural areas of PNG, and unfortunately clinical validation for the same is often lacking. Despite the existence of a large corpus of medical annotation of plants for PNG, previously unknown medical uses of plants can be uncovered. Furthermore, comparisons of medicinal plant utilization is possible if databases are reformatted for consistencies that allow comparisons. A concerted effort in building easily comparable databases could dramatically facilitate ethnopharmacological analysis of the existing plant diversity.


Background
Papua New Guinea (PNG) is a largely rural country characterized by at least 800 ethnic traditions dispersed over 462,840 km 2 [1,2]. Most of the population resides in small villages, situated in diverse environs that range from montane rainforest to lowland river deltas and small tropical islands. Settled 49,000-44,000 years ago (Ivane Valley in the PNG Highlands) [3], PNG is blessed with extraordinary biological diversity and a rich but fragmented cultural tapestry of customs, art, spiritual beliefs and medicinal knowledge.
The East Sepik Province is situated in the northwest of the country bordered by the West Sepik Province (West), Madang Province (East), the Bismark Sea (North) and Enga Province (South). East Sepik (43,426 km 2 ) is characterized by mountainous terrain to the south and west and the costal floodplain of the Sepik river, which flows west to east through the province [4]. The approximately 350,000 inhabitants have to rely on 37 health centers for provisioning health care and heavily supplement western medicines with traditional medicines (TM) [4,5]. The 10 % mortality rate for children under 5 years reflects the difficulty of providing adequate health care in the East Sepik Province. In an effort to supplement health care with effective traditional medicins the University of Papua New Guinea (UPNG), endorsed by the PNG government, struck a collaboration with the WHO to develop reliable traditional medicines (TM) and safe practices (outlined in the 2001-2010 PNG National Health Plan [6]). Part of this project includes traditional medicine surveys performed by UPNG students working in their kinship ("wantok") communities. The data are then recorded in a proprietary database maintained at UPNG [7]. This database serves as central repository for PNG traditional medicine practices, preserving cultural traditions from many diverse communities.

Methods
The TM surveys are performed by UPNG students who are instructed on plant identification, preservation, herbal medicine use, and trained on how to administer the survey instrument entitled "Information sheet on traditional herbal preparations and medicinal plants of Papua New Guinea." The survey questionnaire is the basis for semistructured face-to-face interviews with healers, herbalists, birth attendants, and bone setters. Field vouchers of medicinal plants (twigs with leaves, fruits, flowers, nuts, etc.) are harvested under guidance of the healer and dried and compressed in newspapers. Photographs, descriptions and the pressed plant samples are assigned a voucher number and deposited with the UPNG Herbarium for later identification and reference [8].
The data concerning plant use are written up under supervision into student authored reports and the plant information is entered into the UPNG Traditional Medicines Database, which contains the combined data from reports generated by a decade's work in this endeavor. It is the student reports that provide the base information for this current report.
Four student reports from the East Sepik Province representing four distinct language dialect communities have been compiled here: A compilation of references for medicinal plants described from Papua New Guinea is currently in progress (50 references) in our lab. These references were used to determine if plants collected in the current East Sepik survey work have been previously reported for medicinal use. Comparison of this East Sepik medicinal plant report to our previous reports from Bougainville [8] and the Eastern Highlands [9] was accomplished after editing the previous two reports to match the current format, including codes for conditions treated.
Overall flora distribution data for PNG was obtained for the provinces of East Sepik, Eastern Highlands, and the autonomous region of Bougainville from the PNG Plants Database [10]. The data was imported into Microsoft Excel™, formatted and edited as necessary, then processed with standard Unix (Linux) utilities to produce a formatted list containing the plant family, genus and species (if known). Duplicated instances of plants in the list were removed. The family names were split off, sorted and processed using a Python script on a Raspberry Pi Model B (http://www.raspberrypi.org) to quantify the number of instances of each plant family cited in the list. In general, when multiple names for the same species were found, we attempted to match plant family names to the PNG Plant Database versions to allow for quantitative comparison. Statistical analysis was carried out exactly as previously described by Weckerle et al. [11] using the "beta.inv" function provided in Microsoft Excel™ to calculate the 95 % inferior and superior credible intervals for the data. Comparison of the currently reported East Sepik medical plants to the Traditional Medicines Plant Database maintained at UPNG was carried out similarly. Discrepancies in plant family names were resolved as described above, the family names were adapted to the family names in the UPNG Traditional Medicines Database; resulting in a slightly different number of plant families for the traditional plant uses reported.
Plant families which are considered overused have an inferior credible interval that lies above the superior credible interval for the regional total data. Plant families that are considerend underused have a superior credible interval below the inferior credible interval of the regional total data set distribution.

East Sepik reports
The combined student reports contained 299 entries (including unidentified plants (n = 6)) from the East Sepik province of PNG. The reports collated 205 plants, of which 139 were identified to species and 66 to genus, from a total of 71 families. Three reports were from areas of estimated 150 m to 300 m elevation (DK, BK and GW voucher numbers), DK from Waskuk Hills in the center of Sepik province, and BK and GW from elevated areas of the province north of the Sepik river plain and southwest of the capital Wewak. One report (MS voucher numbers) came from an island (Kairiru Island) situated close to the coast of East Sepik. The combined dataset is presented in Table 1 for all four areas of the East Sepik province under consideration.

Shared and unique plants
We found a number of plants were reported as used in common amongst these areas. Many plants had many overlaps in use, preparation, and disease (Table 1). However, among the plants identified to species level, only four species were reported in every survey: Alstonia scholaris (L.) R.Br., Cassia alata L., Passiflora foetida L., and Zingiber officianale Roscoe. The number of plants unique to one or another of the four reports was surprisingly large in comparison to the previous reports [8,9]. A total of 80 genera, of which 29 are identified to genus level and 51 to species level (see Table 2), were not shared between any of the four study areas.

Plant parts utilization, preparation, administration and diseases treated
In general the areas studied were similar in the relative utilization of plant parts ( Fig. 1) with leaves predominating followed by bark and sap as next most common (with the exception of GW where roots were more commonly utilized than sap). The MS sample set reported a large number of young shoots/young roots stipulated for use in comparison to the other reports, where "young" was not specifically stipulated. The DK and GW reports only cited use of shoots. Only DK reported the medicinal use of nuts.
The method of preparation (Fig. 2) shows a similar pattern amongst the reports: use of succus (expressed juice) was most commonly reported, followed by decoction and direct application of the raw plant material. Usually direct application meant placing the material on a wound or skin ailment after minimal handling. Similarly all reports contain inhalation of smoke or vapor, heat treatment and cooking prior to utilization. DK reported a much higher frequency of cooking the material than the other areas. Boiling as a method of preparation was only mentioned in the MS and GW reports, while mastication (chewing) was reported in all except MS. Typically heating implies later consumption or preparation of steam for inhalation, however, in the GW report heating is a method to prepare the plant material prior to topical application (labelled HR-Heated-Rubbed). Another mode of preparation was mastication and spitting on the affected area. This was relatively common in the MS report and mentioned in the DK report, but not noted in the the other two areas. Only from the DK report is the reverse utilization of the plants reported, where in one instance Homalium foetidum (Roxb.) Benth. was utilized in a reverse-fromexpected manner. In this case, the blood of the patient was placed under the bark of the tree with the expected result being a lessening of knee pain and strengthening of bones as the tree grew. This clearly implies a spiritual/ magical connection of plant and patient.
The routes of administration for plant based medicines reported by DK, BK and MS were about evenly divided between oral or topical routes (Fig. 3). The exception was the administration practices reported by GW where oral consumption outpaced topical application (3:2 ratio). Inhalation was reported only once for the DK and BK areas, and more frequently in the BW and MS areas. The lone outlier for route of administration was from the DK report in which patient material (blood) was transferred to the plant (as described above).
Ailments treated (Fig. 4) with plant based medicines were categorized into 28 groups, sorted according to the target site, in order to to minimize possibly uncertain medical judgements or clinical misdiagnoses. Many described symptoms can likely accurately be ascribed to their appropriate causative diseases, but in the absence of independent clinical confirmation the decision was made to present the data in as unbiased a way as possible. Therefore, the category of "SKIN" contains both               infections (e.g., "Grille") and ectoparasitism (e.g., scabies); the category "REP" contains all sort of reproductive conditions, e.g., impotence, abortion, menstrual syndromes, contraception and fertility, etc. The exception to this method of categorization is malaria, which is generally well recognized throughout the Sepik. Overall, skin conditions were most frequently treated (

Most common families of plants used by healers interviewed
By far the most common genus was Ficus (11), followed by Euphorbia (7), Piper (6) Capsicum annuum L. and Curcuma longa L. are commonly grown in many gardens across PNG, yet it was surprising to note the paucity of medicinal uses previously reported for PNG. Ipomoea pes-caprae (L.) R. Br. (BK 020/06 & MS 26/04) also did not appear to be part of the older literature, however, it was recently found to be used in the New Britain Province where the leaves are rubbed onto the skin affected by jelly fish stings [12]. The sap is used in the BK area for respiratory ailments, and the succus from the leaves is reported by MS to be used in Kairiru for fever/pain via oral consumption.

Comparing East Sepik with Eastern highlands and Bougainville provinces
The combined dataset of the East Sepik, Eastern Highlands and Bougainville reports encompasses 276 plant genera, of which only 22 were reported in common from our other published data sets; Bougainville 112 genera, Eastern highlands 121, and East Sepik 154 genera (see Fig. 5). The frequency of shared genera is given in Table 3. The plant genera with the highest common use citations (> = 10) are Ficus sp. 29, Alpinia sp. 16, Piper sp. 15, Syzygium sp. 12 and Alstonia sp.11. The predominance of Ficus sp. is not surprising since Ficus represents a very large genus in PNG [13].

Regional utilization of plants
Comparison of plants used medicinally in our published data sets to a general list of plants from the same regions allowed for an analysis of utilization preferences. Medicinally over-and under-represented plant families are given in Table 4, while medical plant utilization is given in Table 5. Comparison shows that the number of plant families significantly underutilized, when compared against the regional flora, breaks down as follows: in East Sepik (ES) province Poaceae are underutilized, while in the Eastern Highlands (EH) and Bougainville (BV) Orchidaceae are underutilized.
The number of plants overutilized varies (ES: n = 15; EH: n = 25 and BV: n = 12) but is relatively stable as percentage of plants found in the regional database at 0.66, 0.7 and 0.78 % for ES, EH and BV, respectively. East Traditional inspection of the information gathered yielded information about plants not widely used, poorly annotated or used for different ailments than those in locales where use of the plant is more common. Plants without annotation in the recent PNG Medicinal Plant Literature include: Alocasia cucullata (Lour.) G. Don surprisingly did not yield any crossrefernces in the PNG database, even when using synonyms. It is used in Chinese medicine for snakebite, abscesses, rheumatism, and arthritis [14] and has recently been identified as containing anticancer compounds [15,16].
Averrhoa carambola L. (starfruit) fruit is used for cuts and asthma in PNG, and also widely used throughout the world for a variety of ailments, seemingly only in India as antihemmoraghic [17].
Caryota mitis Lour. has no further medicinal annotation for use in PNG, but is used several Asian countries for a variety of ailments, e.g., against hemorrhoids, male sexual dysfunction, and rheumathoid arthritis in Bangladesh [17].
Chrysopogon aciculatus (Retz). Trin is used in the East Sepik for swelling. The plant is used in Ayurveda as a diuretic [17,18].
Clitoria ternatea L.is used for infertility in PNG and similarly in Ayurveda, where fresh root juice in fresh goat milk is used for pregnancy [18], however, the plant is used for a dizzying array of conditions and ascribed activities [17].
Endospermum medullosum L.S.Sm. has been described previously as used against rheumatism [18], perhaps similar to the use against general body pain in the East Sepik.
Used as a contraceptive in the East Sepik, Erythrina merrilliana Krukoff reveals a dearth of information regarding medicinal uses. The plant is however known to produce toxic alkaloids [19].
Gnetum gnemonoides Brongn. yielded very little information as to medicinal use, but has been described to contain a variety of stilbenes [20].
Hemigraphis reptans (G. Forst.) T. Anderson ex Hemsl. is used in the East Sepik as the whole plant to treat centipede bite. The root is expressed into water to facilitate birth (speeding up delivery) on Vanuatu [21].
No medicinal use annotation was found for Hydriastele costata F.M. Bailey and therefore it may present one of the plants which is used very rarely for that purpose.
Macaranga clavata Warb. is used in East Sepik for skin infections, but has no recent mention in the literature for  Planchonia papuana R. Knuth appears to be not used medicinally elsewhere. It is a timber tree and perhaps as such has not attracted attention; however, in an antiviral screen in our lab fractions from P. papuana exhibited anti-HIV activity [25].
Plectranthus parviflorus Willd., along with Plectranthus blumei (Benth). Launert, and Plectranthus myrianthus Briq. belong to a genus prominent for production of essentials oils [26] and with multiple annotations for antimicrobial activity, but do not seem to be described elsewhere in the PNG plant literature. The utilization of these plants for sores, ulcers and fresh cuts appear to be in line with the activities of chemicals found in Plectranthrus species [27].  Riedelia corallina (K. Schum.) Valeton, in the Zingiber family, is used for menstrual cramps, but seems to be otherwise undescribed for medicinal uses elsewhere.
The leaves of Schismatoglottis calyptrata (Roxb.) Zoll. & Moritzi are used in East Sepik to treat skin sores. No other mention was found in the PNG literature. The stems of Schismatoglottis calyptrata (Roxb.) Zoll. & Moritzi are however used in Chinese medicine for treatment of lumbago and arthralgia [18].
Sterculia shillinglawii F. Muell. has no previous annotation for PNG, but is known to be used in the Solomon island as a tonic and to reduce fever [18].
Tinospora arfakiana Becc. likewise lacks further medicinal descriptions from PNG and does not seem to have been studied from any other area, making it a potentially understudied plant.
Uncaria lanosa var. appendiculata (Benth.) Ridsdale was mentioned twice in the reports and in both instances to treat fever, but also gastrointestinal diseases, malaria, and malnutrition. No other mention for ethnomedical use could be located from PNG or other locales. However, a recent publication hints at a potential anti-depressant effect of ethanolic extracts of Uncaria lanosa var. appendiculata (Benth.) Ridsdale [28].
Uncaria orientalis Guillaumin, used to treat shortness of breath in the East Sepik, lacks pharmacological annotation, but has been investigated extensively biochemically [29,30].

Conclusions
This report shows that in the East Sepik province of PNG the patterns of plant usage for medicinal indications is highly varied. This is true even though many of the same plants are used in ethnologically distinct regions. There is a tendancy for widely used plants to be used for multiple diseases, often with differing preparation of the parts utilized and differing modes of administration. One such example is Alstonia scholaris (L.) R.Br. which shares only the route of administration between all areas. Regardless, plants not previously documented as being used medicinally can still be uncovered, e.g., Cascabela thevetia (L.) Lippold, a plant known to contain highly toxic cardiac glycosides [31] and Dendrocnide cordata (Warb. ex H.J.P. Winkl.) Chew cannot be found as being used medicinally, however, toxicity from leaves, which are used in East Sepik, has been documented [18].
Comparison of plant utilization across study areas can likewise uncover plants which share use. A good example is the genus Alpinia, for which gasteroenterological, respiratory and reproductive use are cited for Bougainville. In the Eastern Highlands it is used for gasteroenterological and respiratory conditions. In the East Sepik it is also usedfor respiratory conditions. Alpinia is in the ginger family, widely used culinarily and medicinally around the world, with traditional medicinal uses for several of the described symptoms.
Likewise, dissemination of knowledge of useful phytomedicinal practices amongst areas that share key flora may aid health practices in those areas. In any case, further studies and phytochemical analyses need to be completed before addition of plants to the pharmacopeia for PNG (a goal of the National Policy for Traditional Medicne in PNG). The UPNG Traditional Medicines Database, while still being populated with data, can already be utilized to show correlations and extract lead information for targeting certain plants for further study. Further enhancements and perhaps adaptation of other data sources (e.g., the PNG Plant Database with up-to-date plant nomenclature) would drive statistical discovery of  [9,25] medicinally neglected plant genera. It is shown here that transregional comparisons are possible, but require careful recoding of previous reports and standardization of database entries and terminology. Analysis of frequency of use of plant families in the medical tradition points to certain biases. This can ultimately be useful in targeting plants for biochemical investigation. However, if the desired outcome of the ethnobotany endeavor is to highlight useful plants for the pharmacopeia, then finer grained data is needed in order to dissect the wealth of information gathered, (e.g. precise geographic location including environmental conditions, etc.). Annotation with biochemical information, conservation status, toxicity data would yield utility for a more diverse set of scientists. To this end the diverse efforts of PNG botany, ethnobotany, ethnopharmacology and plant conservation need to collaborate more rigorously to define useful interfaces for each other's data needs. Nevertheless, we have been able to successfully show that medicinal plant use in terms of families utilized in the East Sepik resembles Bougainville provinces more than it does the Eastern Highlands. Future work with larger data sets will address whether such similarities are due to similarities of available flora or other causes.