Finding Something

Hip Arthroplasty in Osteopetrosis

This is the sort of subject that might come up on you suddenly because patients with osteopetrosis are living longer. Many of them suffer from severe osteoarthritis of the hip and so may come for consideration of an arthroplasty at a young age. The technical problems of doing a hip replacement in such abnormal fragile bone may well be daunting.

The Medline search engine works by taking the words you enter and analyzing them. We will see shortly what exactly they do and how to find out. It means that you must be careful to choose the words which you think will be in the paper you want to find. In the text box below I have entered the words I initially used for this search

osteopetrosis hip arthroplasty

You can copy and paste them to the PubMed dialog box you have already opened. And you can use the box to edit and alter the search string you are using so you don't forget it  (workshop instructions in green)

When you click on the Go button the site will analyze and interpret this search string in quite a complex fashion (see below) then conduct the search of the database.

You can see that 7 papers were returned

The default type of display is the summary which does not give the abstract. As seen below it gives the authors names as a link to the abstract, the title of the paper, the PubMed ID number (PMID) and the status. Indexed for MEDLINE means that the librarians have indexed this paper fully.

If you click on the third citation you will see the following abstract. If you wish

you can click on the link to the International Orthopaedics Journal site which has the abstract and a pdf version of the paper available to subscribers. Many other journals have a Pay Per View system where you can view an individual paper for a modest price.

So this very simple exercise found "something"; in fact 7 papers which mention hip arthroplasty and osteopetrosis. With a small number like that it is worthwhile to read all the abstracts. Even the unpromising looking first paper Benishou et al contains useful stuff -

As many as 81% of our patients experienced clinical manifestations. Fractures were common (78% of patients) and healed slowly. Hip osteoarthritis developed in 27% of patients and required arthroplasty in 9 of the 16 affected hips. Nonmandibular osteomyelitis occurred in 4 cases (11%). Twenty-four percent of patients had thoracic or lumbar scoliosis. Orthopedic surgery was performed in 52.8% of patients, of whom half had at least three surgical procedures for internal fracture fixation, arthroplasty, limb deformity correction, or treatment of surgical complications. There was a high rate of surgical complications including nonunion, infection, prosthesis loosening, and intraoperative fractures.

Clicking on the abstract links to read them causes them to change colour (usually to orange) and this helps distinguish abstracts you have already read. This is a good idea if you are going to extend the search. You can do this by clicking on the Related Articles link for the citation of your choice. You will get inundated with off subject articles. However, it is easy to pick out the ones you have already read because the link has changed colour.

In fact there seem to be very few other reports which (by title) seem to be about arthroplasty. Incidentally you can see why paper 3 in the list above wasn't "caught" originally. The author uses the term endoprosthesis, not arthroplasty.

This search was successful. One entry turned up a small list which was mostly on the subject and certainly offers a consensus view that THR in osteopetrosis is technically demanding but possible and valuable. The idea of surface replacement arthroplasty in this condition was certainly one I hadn't thought of.

Why did the search work? (If you think that we have already accomplished what we set out to do and want to leave, you are right; please go ahead!)

The search string was carefully chosen to include as many elements as possible. These elements were terms that virtually had to be in a paper on the desired subject.

• Osteopetrosis - leaving that out would have opened the floodgates!
• Hip - it would occur in virtually any paper talking about hip arthroplasty and not in ones about knee arthroplasty.
• Arthroplasty - this is probably the commonest term used in the literature but there are synonyms such as joint replacement  (and endoprosthesis!)

PubMed assumes the Boolean logic term AND between any two words in your search string so you might expect it to have translated the search request as Osteopetrosis AND (together with) Hip AND Arthroplasty. It's instructive to see what the site actually did.

To understand this we need to understand the NLM's special vocabulary of medical terms MeSH which stands for Medical Subject Headings. Certain words have been chosen to be the one and only MeSH term for a particular subject. In this case we have actually picked three terms which are in fact part of the MeSH terminology

• Osteopetrosis - this is a MeSH term and synonyms like Albers-Schoenberg Disease are not.
• Hip - pretty difficult for that to be anything but a MeSH term
• Arthroplasty - this is also in the list of MeSH terms in the "tree" of surgical procedures

The end "twig" of this tree is "Arthroplasty,Replacement, Hip" and this is the most specific MeSH term for what we were looking for.

To appreciate the importance of MeSH terms we have to understand what the NLM librarians do when they receive a journal issue. The title and abstract are entered into the database almost at once. Then the staff look at each article and determine what MeSH topics the article covers. The topics chosen are listed in the full citation. The MeSH terms chosen for the Matsuna & Katayama article (PMID: 9498153) were

The whole article was used to make this list and from it we can deduce that there was one male and one female patient; one was an adolescent and the other middle aged; the study included case reports and follow-up and that range of motion was reported or at least an issue.

If you use a MeSH term in your search string or if the PubMed system interprets your search string as a MeSH term then it will pull up an article that has that term. It will do that whether or not you used the exact words that appear in the citation. For example, if you searched for Osteoarthritis/surgery  PMID 9498153 would be somewhere in the list, even though the authors have used osteoarthrosis in the abstract and never mention osteoarthritis. The librarians have indexed it as about  osteoarthritis.

So now let's pull apart the search string we found in "details" and see what the search engine actually looked for

Translation back into English (sort of) -

Articles which either are indexed as being about osteopetrosis or contain the word osteopetrosis and also have been indexed as being about the hip or contain the word hip and in addition have been indexed as being about arthroplasty or contain the word arthoplasty.

I'm sorry! I can't boil it down any further.

Orthopaedic surgeons like to understand how things work. In this case it does help to know that PubMed will take your words and interpret them in ways you might not expect. Remember you can always see what PubMed is doing by looking up the "details"

Now it's time for you to enter your own subject and see you you fare. Go back to the text box at the beginning of this page

Further details on the use of MeSH are in the section on broadening and narrowing the search

Alternatively try one of the other "aims" by clickng on the Navigation bar to the left.

Myles Clough mylesclough@shaw.ca
Clinical Instructor, Department of Orthopaedics, University of British Columbia
B.C. Canada
Feedback, Comments and Questions welcome.