I've been wrong about treating pleural effusion for 25 years
The Study of the Week shocked me. A Danish team (who else) decided to submit thoracentesis for HF-related pleural effusion to the RCT test.
Patients who present with acute fluid accumulation and congestion due to heart failure often develop something called a pleural effusion.
The fluid is not in the lung, but around the lung. It compresses the lung, creating shortness of breath, chest pressure, and other symptoms.
We call such a fluid accumulation extra-vascular because it is not inside the blood vessels. Rather, it is in a separate space outside of the lungs. It’s an important distinction because it led me to believe that diuretics would be less likely to resolve the fluid collection.
For moderate to large effusions, I have recommended something called thoracentesis. We used to do this at the bedside, but now a patient goes to radiology and a doctor uses ultrasound to place a catheter into the fluid collection. The tube drains the fluid out with either gravity or a vacuum bottle.
Patients come back from radiology feeling better. And I would feel better for being so smart.
Well. It turns out there was scant evidence that this procedure provided benefit.
Last week, a Danish team published results of a randomized trial comparing thoracentesis to standard medical treatment for the treatment of patients with heart-failure-related pleural effusion. Medical therapy is mostly diuretic drugs.
TAP IT is an elegant trial. Elegant in its simplicity.
Patients with heart failure and pleural effusion were randomized to either thoracentesis plus medical therapy vs medical therapy alone. Pleural fluid can have other causes besides congestive failure so the team had criteria to identify effusion due to heart failure.
In three Danish hospitals, the research team screened 255 patients to enroll 136 patients. These were elderly (81 years) patients with depressed heart function (LVEF 25%) and modestly high frailty scores.
Patients with massive pleural effusions were excluded. As were patients who had normal heart function.
The primary outcome was days alive out of the hospital. I like that choice because it measures overall health burden. They had many secondary measures, most notably, quality of life assessments at 14 and 90 days.
The average amount of fluid drained was a 1062 cc (similar to previous observational studies).
The results were clear. In short, every outcome measure was similar.
Below is the primary outcome and the quality of life results. KCCQ assesses health status, symptoms, functional limitations, and quality of life in patients with heart failure.
Other outcomes were also not statistically different. Days alive and not hospitalized for heart failure were also not substantially different. No subgroups were found that had benefit from thoracentesis over medical therapy; not even those with large effusions.
There were also no differences in length of stay, diuretic requirements, weight loss, kidney function or status of medical therapy by hospital discharge.
There were more adverse effects in the thoracentesis group. One in four patients in the thoracentesis arm experienced minor complications or discomfort during or after the procedure.
The authors identified a number of limitations, most of which center on the pragmatic nature of the trial. Pragmatic trials attempt to mirror standard clinical practice. For instance, the authors did not have data on completeness of drainage.
The biggest limitation is selection bias. That is: patients most likely to benefit from thoracentesis were not randomized because their doctors wanted them to have drainage. This is always a concern in trials, but I find it reassuring that the number screened (255) was not that much higher than the number randomized (136). I’d be more worried about selection issues if they screened a 1000 patients and only randomized 100.
Comments
I recall a lung doctor telling me a few years ago that thoracentesis in acute heart failure had little supportive data. He was right; but I still believed it was a valuable procedure. No data did not mean no benefit.
My thinking was this: the fluid drained out looks impressive in a jar at the bedside. Your brain sees it and thinks: we really helped this patient getting that fluid out. Gosh that is a lot of fluid!
Also, patients come back feeling better. You think it can’t be placebo effect because of that big jar of fluid. Of course people feel better when you decompress the lung.
The TAP IT trial shows that I WAS FOOLED. I fell for one of Medicine’s biggest mistakes: just because a treatment makes sense does not mean it is beneficial.
Thoracentesis to drain fluid from the pleural space in patients with heart failure has been done for at least 40 years. Yet no one thought to do a proper trial. Until now.
What a beautiful trial. Though I don’t love being wrong, I do love medical reversals. They teach us our most important lesson: staying humble.
I close with an important caveat. Pleural fluid accumulates for many reasons. Heart failure is only one. For instance, it is often valuable to drain fluid to get a diagnosis. And the neutral TAP IT trial does not apply to patients with other conditions.
Congratulations to the Danish team. TAP IT is one of the year’s best trials. Not only because it clarifies treatment of pleural effusion in heart failure (e.g. less is more), but mostly for its lesson in staying humble and requiring evidence for even the most common and sensical treatments.
Thank you for bringing this to my attention. I am impressed that you highlight the fact that your ideas did not hold up. I am so old that I was taught "Never let the sun set on a pleural effusion" (and, yes, I note that this article only pertains to pleural effusions from heart failure and the training maxim refers to every kind of pleural effusion imaginable). However, when I happen upon a pleural effusion in an elderly patient with heart failure who really doesn't want to go out again to the hospital to get it drained, I will stop feeling that I should try harder to get them to get it drained.
As a pulmonologist of over 30 years, I knew this. I am surprised that some may have thought otherwise, but perhaps experienced intuition has some value. The interesting practice that I see repeatedly today, which clearly has no evidence, is that of interventional radiology approaches to pulmonary embolism in hemodynamically stable patients. Similar to thoracentesis for heart failure, many today just can't fathom that removing clot from pulmonary arteries mechanically or with low doses of thrombolytic makes no difference in stable patients. In fact, it may be potentially risky in a relative sense. Perhaps our infatuation with quick results, financial reward for procedures, fascination with new technologies, or just human stubbornness to accept our desires and beliefs over old fashioned evidence is the cause. Interestingly, every study of IR therapy for stable PE patients which demonstrates no benefit ends with the phrase "more studies are needed". The implication is, "well we'll just keep trying until we make the study show what we want it to". Bet on it...