Stufkens SA, Knupp M, Horisberger M, Lampert C, Hintermann B.  Cartilage lesions and the development of osteoarthritis after internal fixation of ankle fractures:  a prospective study.  J Bone Joint Surg Am.  2010 Feb; 92(2): 279-86.   (Pubmed ID#: 20124053)

WHY did the authors undertake this study?
Ankle joint arthritis is generally the result of a traumatic event, and can represent a significant sequela following open reduction-internal fixation (ORIF) of an acute fracture. This may occur as a result of unrecognized damage to the articular cartilage at the initial time of injury. The aim of this study was to examine the relationship between arthroscopically-detected initial cartilage damage and long-term patient outcomes following ORIF of acute ankle fractures.

HOW did they attempt to answer this question?
The primary outcome measures of the study included a radiographic measurement of ankle osteoarthritis (using a modified Kannus score) and a clinical measure of patient function (the AOFAS hindfoot score) at follow-up. These were then correlated with extensively documented operative reports of intra-articular pathology obtained via arthroscopic examination at the time of injury prior to surgical reduction. The grade and exact location of any cartilage defects were recorded.

A cohort of 106 patients with a mean duration of follow-up of 12.9 years (range, 11.3 to 14.8 years) was identified for inclusion in the study.

WHAT were the specific results?
A total of 81% of patients had some form of cartilage damage during the initial arthroscopic examination, including 65% with a lesion on the talus, 50% with a lesion on the tibia, and 39% with a lesion on the fibula.

Statistically significant differences in radiographic (odds ratio 3.4; p=0.04) and clinical (odds ratio 5.0; p-0.02) outcomes were found between patients who had a cartilage lesion versus those who did not.  Deep lesions (accounting for >50% of cartilage thickness) on the anterior talus, lateral talus and medial malleolus were all statistically associated with the development of clinical signs of osteoarthritis.

HOW did the authors interpret these results?
From these results, the authors concluded that the presence, location and severity of intra-articular cartilage damage at the time of injury could help predict poor clinical outcomes and the development of ankle joint osteoarthritis in patients with an ankle fracture.

Decousus H, Quere I, Presles E, Becker F, Barrellier MT, Chanut M, Gillet JL, Guenneguez H, Leandri C, Mismetti P, Pinchot O, Leizorovicz A; POST Study Group.  Ann Intern Med. 2010 Feb 16, 152(4): 218-24. (Pubmed ID#: 20157136)

WHY did the authors undertake this study?
Superficial venous thrombosis (SVT) may not be as benign a prognosis as initially assumed.  The aim of this study was to determine the association of SVT with deep vein thrombosis (DVT) and pulmonary embolism (PE).

HOW did they attempt to answer this question?
There were several outcome measures in this study. First, the prevalence of concurrent DVT and PE was determined in patients initially diagnosed with SVT (defined as a symptomatic subcutaneous noncompressible hypoechoic area in the course of an identified superficial vein). Second, the patients were followed for 3 months following diagnosis of SVT for the development of a thromboembolic complication (including DVT and/or PE).  And finally, an attempt was made to identify specific risk factors for the development of this pathology.

The population cohort consisted of 844 patients enrolled by vein specialists in France.

WHAT were the specific results?
Approximately 25% of patients who were initially diagnosed with SVT were also diagnosed with DVT or PE at the time of diagnosis. An additional 10% of patients developed a thromboembolic complication in the 3 months following initial diagnosis of SVT.  Specific risk factors that were found to be associated with the development of these thromboembolic complications at the 3-month mark included male gender, previous cancer, absence of varicose veins, and history of thromboembolic complication.

HOW did the authors interpret these results
From this data the authors concluded that symptomatic SVT is not an entirely benign diagnosis, and that these patients carry an increased risk for the development of a thromboembolic complication. These patients should be educated about their risk, and followed closely for the development of complications.

Let’s take a closer look at the topic of consecutive patients as it applies to these articles. In the critical analysis section of PJC#12 we talked about taking a look at the “experience of the surgeon”, specifically by examining how many patients were enrolled in a study over what period of time.  One very important word that you should also look out for in this “Material and Methods” section of study when performing a critical analysis is “consecutive”. If the word “consecutive” is found, then we can assume that the authors took a look at all of the patients who met inclusion criteria, instead of simply cherry-picking patients who had a favorable result (or an unfavorable result - depending on what conclusion the authors are looking for!). The presence or absence of this one word can make a big difference with respect to the validity of a study.

In the “Materials and Methods” section of the Stufkens et al article we are told that a “consecutive cohort of 288 patients” were identified that met study inclusion criteria. However, only 106 of these patients ended up in the final statistical analysis. What happened to 63% (182/288) of the patients that met inclusion criteria?  In other words, why were the authors only able to study 37% of the patients that they wanted to? The authors actually do a good job of telling us:  71 refused to participate in the study, 57 died, 37 had moved out of the country, 14 could not be located, 2 were excluded because of a bad surgical reduction, and 1 was excluded because of another significant ankle injury.

We now know that seventy-one patients (25%; 71/288) refused to participate in the study. Why? Did they have such a great result that they didn’t want to be bothered, or did they have such a bad result that they couldn’t stand the sight of their surgeon again! The authors go on to tell us that the main reasons patients refused participation in the study was the “requirement of additional radiographic evaluation, and health and age-related problems that limited the patients’ ability to travel to our hospital.” They go a step further and tell us that no differences in demographic data or original arthroscopic findings were identified between patients who participated and chose not to participate. 

While this is obviously a critique of the article (ideally >80% of patients who meet inclusion criteria are included in the final analysis), at least the authors provided us with specific information about it. As critical readers we are forced to make fewer assumptions about the population cohort in this case than if the authors had said something along the lines of “106 non-consecutive patients were identified” in their Materials section. 

This concept boils down to having as much information as possible, and not having to assume why patients are included or excluded from a study. Each time we are forced to make an assumption about a study (whether good or bad), it should decrease how strongly we can accept the results and conclusions. Each patient that meets inclusion criteria should be accounted for. The Stufkens et al article spelled it out for us within the text of the Materials section, while the Decousus et al article provided a nice flow chart (Figure on page 220 of the article), but both tell us about each and every patient that could possibly be included.