Suction should not be routinely employed.
Caution is required because of the risk of RPO [Re-expansion Pulmonary Oedema].
High-volume low-pressure suction systems are recommended.

"A persistent air leak with or without incomplete re-expansion of the lung is the usual reason for consideration of the use of suction, although there is no evidence for its routine use. It is arbitrarily defined as the continued bubbling of air through a chest drain after 48 h in situ. A retrospective review of 142 cases of pneumothorax found a median time to resolution of 8 days which was not related to the initial size of pneumothorax, but longer for SSP [Secondary Spontaneous Pneumothorax]. A persistent air leak was observed in 43 cases, 30 of which were treated with suction. The theory that underpins the role of suction is that air might be removed from the pleural cavity at a rate that exceeds the egress of air through the breach in the visceral pleura and to subsequently promote healing by apposition of the visceral and parietal pleural layers. It has been suggested that optimal suction should entail pressures of -10 to -20 cm H2O (compared with normal intrapleural pressures of between -3.4 and -8 cm H2O, according to the respiratory cycle), with the capacity to increase the air flow volume to 15-20 l/min. Other forms of suction are not recommended. High-pressure high-volume suction may lead to air stealing, hypoxaemia or the perpetuation of air leaks. Likewise, high-pressure low-volume systems should be avoided. High-volume low-pressure systems such as Vernon-Thompson pumps or wall suction with low pressure adaptors are therefore recommended.

The addition of suction too early after chest drain insertion may precipitate RPO, especially in the case of a PSP [Primary Spontaneous Pneumothorax] that may have been present for more than a few days, and is thought to be due to the additional mechanical stress applied to capillaries that are already ‘leaky’. The clinical manifestations are cough, breathlessness and chest tightness after chest drain insertion. The incidence may be up to 14% (higher in younger patients with a large PSP), although no more than a radiological phenomenon in the majority of cases. Sometimes pulmonary oedema is evident in the contralateral lung. Fatalities have been reported in as many as 20% of 53 cases in one series, so caution should be exercised in this particular group of patients.”

From BTS Pneumothorax Guidelines. References removed for clarity, see original here.
Listed in Cases