So far, I’ve looked at the only two papers in the trauma literature that examine the question of chest tube size for hemothorax. As you may recall, both were woefully underpowered. Finding no difference in a study without enough subjects does not infer that the two interventions have the same results. It simply means that a (much) better study needs to be done.

One of these papers admitted that more work needed to be done, the other did not. And the one that admitted no weaknesses has been quoted by some of the pigtail catheter studies I am reviewing this week as supporting their hypothesis. They are using it as the rationale that even small catheters might work. Hmm, faulty premises?

After reviewing the pigtail for hemothorax literature since the beginning of time, I found exactly two papers that address the issue. And really, it’s just one. The first one published in 2012 was the initial series. The numbers were expanded over the following years by the same authors, and the new data was published in 2017. Of interest, the authors cite their own early paper as supporting the effectiveness of using a pigtail catheter, even though it can’t due to very low numbers Let’s dig in.

This one comes from the group a the University of Arizona in Tucson. They prospectively collected data on pigtail catheter insertions from 2008 to 2014. The outcomes studied included initial drainage output, catheter complications, and failure rate (incomplete drainage requiring another intervention).

Tubes and pigtails were placed by attending physicians or residents. Patient selection was at the discretion of the attending surgeon. The total patient group was analyzed, and then it was split into emergent placement vs non-emergent placement. Here are the factoids:

• A total of 307 chest tubes and 189 pigtails were placed; pigtail usage increased over the study period
• Pigtail catheter patients were older overall, especially in the non-emergent group (this was admitted as selection bias in the paper)
• Initial output was higher in the pigtail group and reached statistical significance in the emergent placement group (500 cc vs 250 cc)
• Pigtail insertion complications trended higher for all patients and in the non-emergent group, but not in the emergent placement group (??)
• Failure rates were not different across the groups

The authors state the their study “clearly demonstrates favorable outcomes in pigtail catheter usage.” But does it?

Bottom line: Once again, this is a completely underpowered study. The pigtail results would need to be 2-3 times better than chest tube results to show any statistical significance. But they are not. So being non-inferior doesn’t mean anything with such small numbers. However, if you properly power a study that shows no differences, then they truly are equivalent. But with the work available to date, you can’t just run out and start using pigtails because they are “as good as” chest tubes. 

There were a few statistically significant differences in this study, but again this is clouded by other design problems. The emergent group had significantly more initial output through the pigtails. This is odd from a fluid dynamics point of view. How do you get more of a thick liquid to drain from a tiny tube? 

One potential explanation is the ability to more accurately measure the initial output in the pigtail group. When a chest tube is inserted, there is frequently some blood loss on the bed which is difficult to estimate. But when a pigtail is inserted there is almost never any leakage. It all comes out through the tube. Could the excess pigtail drainage be accounted for by external loss during chest tube placement?

The real bottom line: There are a grand total of three published papers in the past seven years that have tried to deal with tube size in traumatic hemothorax. All of them are completely underpowered and rely on the lack of significant differences to tout that they are equivalent. The real answer is: we don’t know. This is certainly not the quality of data you want to use to change your practice. We don’t know for sure if smaller tubes and pigtails result in more retained hemothoraces or followup procedures. So buyer beware! If you choose to use small tubes or pigtails in your patients, you are in uncharted territory. The first author of the 2012 small tube paper even stated that a larger multi-center is needed. I completely agree! Meanwhile, I’ll stick to big (36 Fr) and bigger (40 Fr) for hemothorax.

Reference: A Prospective Study of 7-Year Experience Using Percutaneous 14-French Pigtail Catheters for Traumatic Hemothorax / Hemopneumothorax at a Level-1 Trauma Center: Size Still Does Not Matter. World J Surg 42(1):107-113, 2012.

Over the next few days I will be reviewing a number of papers that try to determine whether the dogma that bigger chest tube size is better is actually true.

Here are the questions that need to be answered when reading each one to determine if it’s worth its weight:

• How good is the study design? Obviously, prospective is better than retro. How did the authors decide to put in a small vs a large tube? Were there enough subjects to achieve any meaningful statistical significance?
• Were the tubes used actually different? If the small bore tubes are 30 – 32 French and the large tubes are 36 – 40 French, would that make a difference?
• What were the outcomes studied? Mortality and complications like pneumonia and empyema are too crude and uncommon to detect a difference. But what about incidence of retained hemothorax, accidental removal, subjective pain, or clotting?
• Did the authors identify and acknowledge any limitations in their study?
• Do the conclusions match up with the actual results?

Let’s kick off the chest tube size debate with an oldie but goodie. The first paper I’ll review was published back in 2012 by a busy LA trauma center.  They performed a prospective, observational study of their experience with two tube size ranges inserted for hemo- and pneumo-thorax over a three year period. The size ranges were 28-32 for small and 36-40 for large. The size selected was based on the discretion of the attending physician.

A total of 353 chest tubes were placed during the study period. This analysis will only dissect the 275 that were inserted for hemothorax.

Here are the factoids:

• Pertinent demographics were identical for the large and small bore tube patients
• Pneumonia occurred in about 5% of both groups, and empyema in about 5% of both
• Retained hemothorax occurred in 12% of small tubes and 11% of large tubes
• Duration of tube placement was about 6 days in each
• Additional procedures such as thrombolysis, additional chest tubes, VATS, or thoracotomy were 3-6% in both groups and were not statistically different
• Pain scores could only be performed on about 45% of patients, and were not different between the two groups

The authors concluded that there were no differences in complications, tube reinsertion, or need for invasive procedures based on tube size. They also concluded that choice of tube size did not impact outcomes.

Bottom line: The authors seem to be saying that the choice of tube size is not important. And if you only read the abstract or conclusions of this study, you might actually believe it. But wait, the authors end the paper with this telltale sentence:

“Further evaluation of percutaneously placed drainage systems is warranted”

This is code for: “this paper isn’t very good and shouldn’t change your practice; it needs further verification.”

So what are the issues?

• There is huge potential for selection bias since the choice of tube size was based on personal preference. For example, the attending could look at the chest x-ray, see a lot of blood, and decide to use a big tube in that patient. No guidelines or randomization were used.
• The authors did not acknowledge any limitations of the study in their discussion.
• The only outcomes that really counted in this study were incidence of retained hemothorax (which was not very well defined) and additional procedures required. However, if you take the incidence of retained hemothorax in the large bore tube patients and do an analysis of the statistical power of the study, you run into a major problem. Given the number of patients in each of the two groups, this study would only be able to show statistical significance if the number of retained hemothoraces in the small chest tube group doubled! Anything short of 25% retained hemothorax in the small tube group would not be significant. Thus, the authors’ findings that there was no difference between the groups was entirely expected based on sample size. 

So this paper does not really say that there is no difference in using a small vs a large chest tube. It says that it was not sufficiently powered to detect anything but a massive difference. Many more patients (thousands) were needed to answer the question.

So the question remains, does (chest tube) size matter? More in the next post.

Reference: Does size matter? A prospective analysis of 28-32 versus 36-40 French chest tube size in trauma. J Trauma 72(2):422-427, 2012.