PLOS ONE has issued a fascinating expression of concern about data collection in a paper it published late last year on the possible spread of deadly viruses among Indonesian orangutans. The case has been brought to the attention of the Indonesian government, but more on that in a moment.
The article, published last July by an international group of primate scientists led by Chairul Nidom, a virologist at Indonesia’s Airlangga University, sounded an alarm about “wild” orangutans in Borneo: Blood tests of 353 “healthy” animals showed antibodies for viruses akin to Ebola. What’s more, the filoviruses viruses to which the antibodies responded, as New Scientist and other outlets reported when the original paper came out, included strains not previously seen outside Africa (as well as Marburg, another deadly infection).
The article immediately prompted two comments. The first, by a poster called orangutanborneo, raised questions about the scale and logistics of the project:
Would you provide further information please?
1. Where do you captured the wild orang-utans? Can you described the locations etc?
2. Where do you quarantine them?
3. How many people do you employed to monitor the 353 orangutans for around 6 months each?
That was followed by a post from Vincent Nijman, an Oxford University researcher who studies various aspects of conservation and the wildlife trade, including primatology and ornithology. Nijman noted that the finding of Ebola-like infections among wild orangutans would indeed be important. However, he wrote, that’s not what the paper demonstrated — at least, not conclusively:
As the authors rightly indicate it is very significant that these results derive from wild orangutans as opposed to for instance orangutans in zoos, rescue centres or that are kept as pets (although in Indonesia the majority of these are also wild-caught). The latter have been in contact with humans, and possibly domesticated animals or in the case of zoo-animals, exotic animals. The information the authors present in their paper on how they obtained the samples is very minimal, to say the least:
Samples were collected from 353 wild-caught orangutans in Kutai Kartanegara and Palangka Raya between December 2005 and December 2006
Orangutans were kept singly in a cage for an unspecified quarantine period during which their behaviour was monitored and their health was checked.
Finally the healthy and socially-normal ones were released into the forest and were monitored for some 6 months.
The scale of the operation (catching 353 orangutans, keeping them in quarantine during which time they are kept under observation, releasing them into the forest, and monitoring them for 6 months) is monumental and unprecedented.
It is also perhaps significant to note that two of the largest orangutan rescue centres in Indonesia, with hundreds of orangutans in their care in 2006, are located in the two regencies where the authors obtained their samples, i.e. Wanariset Samboja (Kutai Kartanegara) and Nyaru Menteng (Palangka Raya). These orangutans originate from various parts of Kalimantan, have different backgrounds (ex-pets, zoo animals, animals rescued from e.g. plantations) and are often housed in social groups. Is it possible that the orangutans used in the study were in fact ones from these rescue centres?
With the previous commentor I agree that it would be good if the authors could provide much more information on how the samples were obtained that were used in this study. While they do so it would be good if the authors included a statement clarifying whether or not their samples were in fact derived from orangutans in the above-mentioned rescue centres.
Here’s what the paper has to say about the data collection:
Three hundred fifty-three serum samples (from 201 males and 152 females) were collected from wild-caught healthy orangutans (Pongo pygmaeus) in East (Kutai Kartanegara) and Central (Palangka Raya) Kalimantan provinces in Indonesia from December 2005 to December 2006 (Table S1). All samples were collected originally for serological diagnosis of influenza and/or mycobacteriosis and tested in the ABSL-3 facility belonging to Airlangga University, Surabaya, Indonesia. The Indonesian government, for conservation strategies, conducts a regular monitoring of infectious diseases in orangutan populations. Under the direction of the Ministry of Forestry, Indonesia, orangutans were carefully captured by at least 4 people using nets to investigate their health conditions. All captured orangutans were registered to give their names. Animals were anesthetized by intramuscular injection of Ketamine and Xylaxin, and blood samples were taken from the brachial vein. After taking the samples, each animal were kept in a single cage. After quarantine (e.g., tuberculosis, hepatitis B, and so on), they were released to the forest if they showed negative results for the diseases and normal social behavior. All the captured orangutans released to the forest were monitored at least around six months. Part of the previous collection was used for this study. Animal works were performed under the approval of Airlangga University Ethical Committee.
In an interview today, Nijman told us that the average reader might easily infer that the animals involved were all captured in the wild. But that’s highly unlikely:
In 2005-2006 I was doing work on trade of orangutans. I was in Borneo. I knew that this didn’t happen as they described it.
Rather, Nijman said, the animals had been kept in rescue centers — in groups of up one case, nearly 1,000 were on-hand when he saw it, he said, although the number is certainly less now. Many had come through the wild animal markets, having been kept in unhygienic enclosures that are “the perfect storm for virus transmission.” In other words, they weren’t exactly the poster children, er, primates, for the transmission of viruses in their natural habitats.
Nijman said he followed up repeatedly with the editors of the journal to reiterate his problems with the study, persistence that finally paid off with this expression of concern:
Following the publication of this article, the editors were alerted to concerns over the accuracy and clarity of information reported regarding the source of the samples included in the study.
The editors were contacted by representatives of the Borneo Orangutan Society who indicated that the samples included in this study were collected by the society for surveillance purposes as per the requirements under Indonesian law, and submitted to Airlangga University for routine tuberculosis tests. The Borneo Orangutan Society indicated that the samples originate from animals kept in rehabilitation camps and that they had no knowledge of the additional tests carried out for the purposes of this research.
We have followed up with the authors in relation to these concerns and while we have received confirmation that the samples were collected for surveillance purposes, we have not received a satisfactory response to the request for specific details for the origin of each of the samples included in the analyses or the methodological steps involved in their collection.
In the light of this, the editors are issuing this Expression of Concern to make readers aware of the concerns over the origin as well as the methodology followed for the collection of the samples tested in the study. The editors have also raised the concerns in relation to this matter to the attention of representatives of Airlangga University and the Indonesian government.
Unfortunately, what the notice doesn’t say is who in the Indonesian government the journal contacted. According to Nijman, he was told the editors brought the matter to the attention of a “representative of the Indonesian Ministry of Forestry” — which isn’t known for being a hotbed of good governance.
In any case, Nijman said the paper deserves to be retracted.
If you draw you conclusions that Ebola occurs in wild orangutans, for me that is something that is really different from animals in a rescue center that have all kinds of backgrounds.
What’s more, he added:
All they find is unspecific reactions to the antibodies of Ebola.
We attempted to contact Tulane researcher Daniel Bausch, who edited the article, to find out why the journal didn’t retract the paper. And we have emailed Nidom and the other corresponding author, Eri Nakayama, of Hokkaido University in Japan, for comment and will update this post if we hear from them.
Hokkaido might have more than a passing interest in the paper. According to the statement of support:
Japan Initiative for Global Research Network on Infectious Diseases (J-GRID) and the Global COE Program from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), and a Grant-in-Aid for JSPS Fellows Japan. The work was further supported by a Grant-in-aid from the Ministry of Health, Labor and Welfare Japan, and by Ministry of Education and Culture, Indonesia
PLOS ONE has thousands of Academic Editors who handle manuscripts (soliciting reviews, making decisions). These are all volunteer scientists with their own careers and interests who agree to handles a few manuscripts per year to help the research community. Once published, however, concerns seem to be handled by the small staff of paid professional editors. I’m not aware of a policy that suggests that the AEr would even have been contacted, or that the AE would have any special say in determining the response to allegations of misconduct. It would be interesting indeed to hear from Dr. Bausch to find out what role, if any, he has had in the process.
I do worry about PLoS One and whether they really have the expertise available for good reviews of such a wide variety of topics.
They try to match editors to papers, and editors really shouldn’t agree to take papers they don’t have the knowledge to evaluate. At least they give the editor’s name so you can see for yourself. (And of course, the editors usually solicit 2-3 reviewers, so there’s that as well.)
I have noticed that sometimes PLoS publishes papers in my area that I actually know have been rejected by several other journals that have more expertise, and sometimes PLoS publishes papers in my area that I think are really not very good at all. So I wonder if they are really getting the best reviewers and editors for everything.
agreed about PLoS ONE. I had commented earlier about the academic editors of this journal. Some are not qualified. Very strange journal but probably that is the way they want to function.
I’m curious if anyone with an epidemiology background has looked at the statistics section of the paper. It may be an entirely appropriate method, for all I know. It just struck me as odd.
Hi Toby, looks like it’s (semi) standard for the setting of ref. intervals in serology, eg
http://www.ncbi.nlm.nih.gov/pubmed/19738362
but I am guessing it is crucial to have independent evidence (PCR?) that the animals were pathogen-free, otherwise who knows what’s being measured?
That said, the distribution of serological data is generally *very* far from normal, so the entire basis of the Smirnov-Grubbs test might be pretty shaky here, unless the data was (log or otherwise) transformed, which isn’t mentioned in the methods….
This is all bit outside my area, so all the usual caveats apply.