He was a trim 20-something with a minor chest wall injury. He had been surfing and taken a hard pectoral-first landing into the break, the type that reverberates for a beat or two. The monitor looked reassuring, but as the physician entered the room, the family closed in, and the situation became vastly more complicated.
“He has Ehlers-Danlos,” said his mother.
“The vascular type,” said his father.
“He is at risk for aortic injury, and they told us that if he has chest pain, he needs a CT,” said his mother.
“We are worried,” said his sister.
Off to CT he went, and everyone was surprised when the results came back. The good news: no aortic injury. The bad news: a 30% pneumothorax on the left without a lung contusion or rib fracture.
The family was concerned: Was this life-threatening?
“It is not. Why don't we try some oxygen?” the doctor responded and snuck out to ponder his own questions—was this a traumatic or spontaneous pneumothorax? Was conservative management an option here?
The Evidence
The answer is yes and no. On one hand, a just-released systematic review on the topic did not even consider conservative therapy as a treatment option. (Ann Emerg Med. 2020;76[1]:88.) On the other hand, some freshly harvested evidence from Australia and New Zealand suggested otherwise. That open-label, 39-center noninferiority study, the Primary Spontaneous Pneumothorax trial, included patients 14 to 50 with a first-known, unilateral, moderate-to-large primary spontaneous pneumothorax (32% or greater via the Collins method). (N Engl J Med. 2020;382[5]:405; AJR Am J Roentgenol. 1995;165[5]:1127; https://bit.ly/38QB4UC.)
The trial compared immediate intervention (a small 12Fg Seldinger-style chest tube) with conservative management (no immediate invasive intervention and a minimum of four hours of observation), and assessed outcomes (lung re-expansion and utilization and safety outcomes) for 12 months. The noninferiority method of comparison allowed the conservative approach to be billed as noninferior as long as it was not significantly worse than the chest tube approach within a prespecified margin. The maximal allowable difference in eight-week re-expansion rates was set at 9.0 absolute percentage points, using the far end of the 95% confidence interval.
A total of 316 patients were randomized, with 154 patients in the intervention group. Twenty-five patients (15.4%) in the conservative management group failed the noninvasive approach due to symptom progression or physiological instability. At eight weeks, re-expansion occurred in 98.5 percent of patients who underwent interventional management and in 94 percent of those in the conservative management group (risk difference: -4.1 percentage points; 95% CI, -8.6 to 0.5; P=0.02 for noninferiority). Note that -8.6 percentage points fell just inside the -9.0 threshold.
This analysis, however, excluded 50 patients (15.8%) for whom complete follow-up data were not available. So the authors ran a sensitivity analysis in which all missing data were imputed as presumed treatment failures (a hypothetical worst-case scenario), which adjusted the re-expansion “estimates” to 93.5 percent for the intervention group and 82.5 percent for the conservative group, a difference of 11.0 percentage points, which was no longer within the preset noninferiority margin.
The authors described their findings about the primary outcome as “statistically fragile.” Digging deeper, however, one recognizes that this fragility does not extend to the secondary outcome analysis, which showed a clearly favorable signal for noninvasive management. Adverse events (26.6% v. 8.0%) and serious adverse events (12.3% v. 3.7%) were much higher in the intervention group, as were utilization outcomes. Intervention group patients on average received more chest x-rays (10.9 vs 6.4), and were more likely to get at least one CT (19.2% v. 7.8%). They were also more likely to spend time in the hospital during the first eight weeks and to have extended time away from work.
The Trial
Back to our patient. He seemed comfortable, even if slightly channeling familial anxiety. Given the lack of any other traumatic sequalae, the physician concluded that this was most likely a spontaneous, not traumatic, pneumothorax, and the patient was given supplemental oxygen, a touch of Ativan, and ED observation. A couple of hours later, he looked and felt fine. Unfortunately, his x-ray disagreed, and it was progressing in the wrong direction, closer to 35% down now.
Undeterred, the patient and family opted for another round of observation. Again, the two-hour imaging was disappointing. Physicians in Australia or New Zealand might have felt comfortable discharging this patient without an intervention, but with only “statistically fragile” evidence supporting the approach, it did not seem wise to empiricize it with this psychologically fragile family dynamic.
The next step was the True Close device, always intimidating and even more so on a thin patient with a left-sided pneumothorax. As it turned out, cutaneous hyperextensibility, often found in Ehlers-Danlos, required a repeat scalpel poke to facilitate entry and a somewhat protracted access process, which encompassed a pause during which his mother slumped in her chair with a vasovagal event. Ultimately, the vent went in, and the red-colored diaphragm flapped reassuringly. The young man went home with an uncomplicated follow-up course. The physician, though, was left wondering whether he could have spared the man the trauma of the intervention with the conservative approach. Maybe next time.
Dr. Vinsonis an emergency physician at Kaiser Permanente Sacramento Medical Center, a chair of the KP CREST (Clinical Research on Emergency Services and Treatment) Network, and an adjunct investigator at the Kaiser Permanente Division of Research. He also hosts Lit Bits, a blog that follows the medical literature athttp://drvinsonlitbits.blogspot.com. Dr. Ballardis an emergency physician at San Rafael Kaiser, a chair of the KP CREST Network, and the medical director for Marin County Emergency Medical Services. He is also the creator of the Medically Clear podcast on iTunes. Read his past articles athttp://bit.ly/EMN-MedClear.
