Introduction

Pneumothorax refers to the accumulation of air within the pleural space and is traditionally categorised as spontaneous and traumatic subtypes. Spontaneous pneumothoraces occur without preceding trauma and is further classified as primary if there are no underlying lung diseases or secondary when it complicates underlying lung pathologies. Traumatic pneumothoraces can complicate injury (direct or indirect) to the chest and include iatrogenic causes (e.g. after diagnostic and therapeutic interventions).

Spontaneous pneumothorax is a significant global health problem. Historical data from the USA between 1950 and 1974 show that the age-adjusted annual incidence of primary spontaneous pneumothoraces (PSP) is 74 per million males and 12 per million women and for secondary (SSP) cases 63 and 20 per million, respectively [1]. Over the past few decades, the overall incidence appeared to have risen sharply, especially in SSP, which has become at least as significant a problem as primary pneumothoraces. The combined hospital admission rate for PSP and SSP was 167 per million men and 58 per million women a year in the UK from 1991 to 1994 [2].

The overall annual incidence of patients consulting primary care with a diagnosis of pneumothorax was 240 per million men and 98 per million women in the UK, based on records of a general practice database from 1991 to 1995. Data from the UK Health and Social Care Information Centre in 2013 estimated this to be 8000 admissions each year for pneumothoraces which accounted for 50,000 bed days and £13.65 million in costs for the National Health Service [3]. In the USA, the annual cost of spontaneous pneumothorax has been estimated at $130 million [4].

An estimated 15,000 new cases of SSP occur annually in the USA especially in men over the age of 75 who have the highest risk at 600 cases/million/year [2]. Chronic obstructive pulmonary disease (COPD) is the most common underlying disease in patients with SSP; although a vast range of lung diseases are etiologically related.

The growing incidence of COPD means that associated SSP is rapidly becoming a major global issue. The resultant breathlessness can be alarming, and the pneumothorax can progress rapidly; hospitalisation is often necessary. The probability of pneumothorax increases as COPD worsens and those with a FEV1 of less than a litre are at greater risk [5]. The risk of recurrence is as high as 50% following a SSP and measures to prevent recurrences are recommended even after the first episode. Pneumothorax from COPD is a marker of mortality; the chance of dying rises by nearly fourfold with each additional recurrence of pneumothorax [6].

The risk of a traumatic pneumothorax depends on the severity of the injury and has been reported to exceed 35% in some series [7]. Iatrogenic pneumothoraces are probably more common than PSP and SSP combined [8]. Transthoracic needle aspiration is the leading cause of iatrogenic pneumothorax, which occurs in as many as 1:4 patients; ∼10% require chest tube drainage [8]. Mechanical ventilation also induced iatrogenic pneumothoraces in 4% of 553 patients in one series [9]. Thoracentesis, pleural and transbronchial lung biopsies, catheterization for central venous access and cardiopulmonary resuscitation are other recognised causes of iatrogenic pneumothorax.

Management of pneumothorax focuses on two goals: control of any active air leak and prevention of recurrences. Treatment of pneumothorax ranges from observation, simple aspiration, tube thoracostomy to more invasive surgical approaches which may include one or more techniques such as talc poudrage, bleb resection or bullectomy, pleural abrasion and pleurectomy [4, 10]. There is extensive variation in practice between individual clinicians and geographically depending on resources available.

The first step of the care of patients presenting with an acute pneumothorax is to establish if the air leak from the lung has stopped. An active air leak exists if the pneumothorax worsens radiographically after observation or aspiration, or if air is observed to bubble via an inserted chest tube. Surgical intervention may be required, but there is no consensus on the optimal time to intervene [11]. The British Thoracic Society guidelines recommended that a surgical consult is sought if the air leak persists after 48 h [10]. Surgery has risks, is costly and is not readily available in all centres. Post-operative complications and recovery time are also concerns [12]. The risks are particularly significant in patients with SSP (e.g. those with underlying COPD) because of poor lung function and, often, an older age and comorbidities [13, 14].

Considerable efforts have therefore been spent on the search for less invasive alternatives for patients with an ongoing air leak who are unfit for surgery. However, there is a paucity of high-quality clinical evidence to support these options at present. Endobronchial valves have been used in small published series, but success rates vary, and multiple valves are needed because of collateral ventilation [15, 16]. The use of a one-way (e.g. Heimlich) valve attached to a chest tube can help facilitate early discharge and ambulatory management of patients with an ongoing air leak; however, it does not necessarily promote healing of the air leak or pleurodesis.

The exact number of patients who have a ‘prolonged’ air leak is not known. Incidence rates as high as 26% have been reported [17]. In the surgical literature, prolonged air leak is defined as one lasting beyond 5 days following an operation [18]. Prolonged air leak results in longer periods of chest tube drainage and immobility with an increased risk of complications such as respiratory tract and pleural infection [19]. It is the principal reason for delay in hospital discharge [19, 20].

Autologous blood patch (ABP), as a technique for managing air leak, was first reported in 1987 [21]. Many case report and series, as well as retrospective cohort studies, have described the application of an ABP in patients with persistent air leak. Many boast impressive success rates, cost effectiveness and safety profiles. Chambers et al. [22] reviewed ten studies and found an overall success rate of 93% in post-operative patients with prolonged air leak and 92% for spontaneous pneumothoraces.

The instillation of autologous blood into the pleural space is simple, inexpensive, readily available and can be applied to all patients including those with active air leaks. It has few contraindications. If its efficacy can be confirmed, ABP has the potential to benefit a huge number of patients worldwide.

However, a lot of questions remain about ABP. Its mechanism of action is unclear. One hypothesis is that the site(s) of air leak from the lung is sealed directly by clot formed. Others have suggested that blood can induce pleural inflammation and subsequent fibrosis (pleurodesis) by promoting the fibrogenic pathway. Another theory proposes that the clot itself may bring together the visceral and parietal pleural surfaces and create pleural symphysis [23]. None of these hypotheses has been validated.

There is no standardised method of delivery for an ABP. Studies varied in the amount of blood instilled (from 50 to 250 ml) and the dose administered (from one instillation to multiple, usually up to three). In a randomised, single-blinded crossover study [24••], instillation of 1 ml/kg of blood appeared the optimal dose. However, the study was small (n = 44) and only involved patients with advanced COPD.

The blood is drawn from the patient and aseptically injected directly into the chest drain, followed by a saline flush to minimise clotting within the chest tube. The chest tube line is elevated to 60 cm above the patient and is kept unclamped for up to 2 h [21,27,28,29,, 25•, 26–30]. A positional change for 2 h is quoted in many studies [19,28,32,33,34,35,36,37,38,39,, 25•, 27–29, 31–40] though it is debatable if this is necessary, as blood tends to coagulate quickly.

The timing of administrating an ABP after the initial presentation of the patient also differed among studies or even among the patients within the same study. Studies have applied ABP to different types of pneumothorax and results could not necessarily be generalised to other subtypes. Most of the studies were small and uncontrolled. Given most pneumothoraces may heal with prolonged drainage, the lack of a control group makes the data challenging to interpret.

Instilling one’s blood is theoretically safer than injecting a chemical compound. However, concerns remain that blood may promote infection and fibrothorax. Data on the safety of ABP pleurodesis is needed.

Despite these uncertainties, a genuine interest in ABP for pneumothorax persists and the number of articles on the subject continues to grow with contributions originating from various regions of the world (Japan [28, 35], Turkey [25•, 29, 37, 38], Brazil [41], UK [40, 42•], Italy [33, 39, 42•], Spain [31, 32], China [24••], Greece [34] and Germany [36]).

We set out to systematically review published literature to assess the available evidence for the use of ABP in patients with a pneumothorax.

Methods

A literature search was performed using MEDLINE, PubMed and Cochrane databases to identify studies published from January 1960 to June 2016 that reported outcomes in adult patients undergoing blood pleurodesis or blood patch for pneumothorax. The following search terms blood pleurodesis, blood patch, pneumothorax and spontaneous pneumothorax were used. We also manually searched the reference lists of included studies and relevant reviews for additional studies that were not detected by the electronic literature searches. Articles were excluded if they were not published in English, involved ABP for non-pneumothorax patients (e.g. malignant pleural effusion), or if the publication was a review article or editorial.

Out of the 44 articles found, 18 were deemed suitable, which included case series, retrospective studies, prospective studies and randomised controlled trials (Fig. 1). The quality of the studies was assessed according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach [43].

Evidence

The studies included have been summarised in Table 1.

Table 1 Summary of evidence
Full size table

Primary Spontaneous Pneumothorax

Use of blood patch for PSP has seldom been reported. There are no studies exclusively evaluating the efficacy of ABP in PSP. Therefore, it is difficult to elucidate the true success rate of this technique in PSP.

One prospective study included 32 patients (out of 167) with their first episode of spontaneous pneumothorax who developed an air leak for over 48 h [38]. Eighteen of these patients had PSP. It reported that air leak stopped in 84% (27/32) of patients within 72 h of administration of an ABP. There was no recurrence of the pneumothorax after a follow-up ranging from 12 to 48 months. The length of hospital stay did not differ between those who received an ABP and those who did not. Five patients with PSP were retrospectively evaluated after having one to two sessions of ABP application for prolonged air leak. All five were successfully treated with the air leak ceasing within 48 h of instillation.

Cobanoglu and colleagues [29] prospectively compared ABP to talc and tetracycline pleurodesis. Fifty patients (19 patients had PSP) with spontaneous pneumothorax were divided into three groups (20 received ABP, 11 received tetracycline and 19 received talc). Sixty-two percent of patients had an air leak only on expiration and 24% had it on forced expiration. Twenty patients received 50 ml of autologous blood once only. ABP was successful in 75% (15/20) of patients with the majority (19/20) of patients’ air leaks ceasing within 48 h. The success rates for talc and tetracycline pleurodesis was 84.2 and 63.6%, respectively.

Secondary Spontaneous Pneumothorax

Use of ABP has been reported, often as case reports and retrospective audits, in SSP from various lung diseases including mechanically ventilated patients with ARDS [32], patients with interstitial lung disease [35], severe emphysema [24••, 25•, 34], pulmonary tuberculosis [25•, 28] and bronchiolitis obliterans following stem cell transplantation [44]. Again, the reports presented generally encouraging success rates.

A retrospective study reported data from 31 patients with a first episode of SSP and ongoing air leak after 3 days who received ABP [25•]. The air leak resolved in 94% of patients within 24 h and facilitated removal of the chest tube. Pneumothorax recurred in 13% of patients after a mean time of 19.7 months. Four patients underwent surgery because of recurrence (n = 3) and persistent air leak (n = 1) despite ABP after a mean follow-up period of 29 months.

A small prospective observational study of 11 patients with ongoing air leak from a SSP also supported the use of ABP [28]. Their underlying chronic lung conditions included emphysema, tuberculosis, lung cancer and histiocytosis X. Interestingly, ABP was effective in four out of seven cases of persistent air leak in the lungs that were not fully reinflated. Two patients SSP recurred (after 8 days and 5 months, respectively); they were successfully treated with repeat ABP application.

A retrospective audit of 59 episodes of SSP from 34 patients with interstitial lung disease found that ABP was used in 22 episodes and resolved the air leak in 73% of cases [35]. In another study, ABP shortened the stay in intensive care unit by 9 days (adjusted OR = 0.24), reduced weaning time off the ventilator as well as mortality (3.7 vs. 29.6%, adjusted OR = 0.26) [32].

Prolonged Air Leak Post-surgical Lobectomy

A large volume of the literature on ABP centred around its use in prolonged air leaks after lung resection. This included two randomised studies. Shackcloth et al. [42•] performed a single-centre randomised crossover study comparing ABP with standard care (tube thoracostomy) in patients with prolonged air leak post-lobectomy. Twenty patients were randomised using sealed envelopes to study group who received 120 ml of autologous blood on the fifth post-operative day or to the control group who had tube thoracostomy alone until cessation of air leak. ABP was attempted in the control group if the air leak continued at day 10 post-operatively. Patients’ characteristics were similar in both groups. The majority of patients had a small continuous air leak on gentle respiration. ABP was successful in sealing air leak by the next day in 58.6% (17/29). It was effective in sealing the air leak sooner (median 5 vs. 11 days), reducing the duration of chest tube drainage (median 6.5 vs. 12 days) and shortening hospital stay (median 8 vs. 13.5 days). Adverse events were rare, and only one patient developed an empyema.

Andreetti et al. [33] randomly assigned 25 post-lobectomy patients to receive an intrapleural instillation of either 50 or 100mls of autologous blood. The length of air leak before ABP was not specified. Cessation of air leak occurred faster in the group that received 100 ml of blood over those who had 50 ml (1.5 vs. 2.3 days, respectively, p = 0.005).

There are a few retrospective and prospective case series supporting the efficacy of ABP. A prospective series applied ABP to 11 post-lobectomy patients who had a persistent air leak at 7 days. ABP was successful in 10 patients at 48 h, and there were no recurrences after 3 months of follow-up [40]. Athanassiadi et al. [36] evaluated ABP in 20 patients with persistent air leak; 19 patients had resolution of their air leak within 48 h with no complications or recurrence observed during follow-up (between 2 and 24 months). A retrospective study from Brazil showed that ABP was successful in 85% of the cases (n = 27) with a mean time to resolution of 1.5 days [41]. Adverse events were minimal with one patient having an empyema and another fever.

ABP pleurodesis has also been found effective when applied to patients (n = 7) with a prolonged air leak after resection of a pulmonary hydatid cyst [37].

One prospective study combined ABP pleurodesis with induction of a pneumoperitoneum in 39 patients with a residual air space and ongoing air leak after major thoracic surgery [42•]. The air leak stopped in all patients without any complications. However, ABP has been used independently in patients with residual air space with good results [37].

Post-radiological Guided Lung Biopsy

Iatrogenic pneumothorax following imaging-guided diagnostic and/or therapeutic manoeuvres is a major and increasing problem. No studies have specifically investigated the role of ABP for pneumothorax caused by imaging-guided lung biopsy.

Instead, ABP has been used as a prophylactic measure, first by Bourgouin et al. [46] in 1988. This study randomised 140 consecutive patients into two groups based on whether the patient’s identification number ended with an odd or even number. The intervention group (n = 52) had 5 to 10 ml of clotted blood injected via the introducer needle before withdrawing. In the control group (n = 88), the introducer needle was withdrawn at the completion of the biopsy as was usual practice. No statistically significant difference was seen in the pneumothorax and chest tube insertion rates between the groups (28.8 vs. 34.1 and 7.7 vs. 9.1%, respectively).

In a retrospective study of 463 lung biopsies performed over 4 years, 9.7% (45/463) resulted in a significant pneumothorax, of which 4.1% required chest tube placement [47]. The procedure was performed with up to 15 ml of autologous blood inserted through either an introducer needle or a 5-French Yueh Centesis catheter after completely aspirating the pneumothorax. The catheter or the introducer needle was then removed and the patient discharged if medically stable after 4 h of observation. Of the 22 patients who had a blood patch, three (13.6%) required chest tube placement and hospital admission.

The reduction in the need for chest tube placement after an iatrogenic pneumothoraces was further assessed in a randomised controlled trial by Malone et al. [48]. Their study involved 242 patients who underwent percutaneous biopsy of lung or a mediastinal lesion. The blood patch group (n = 123) had 4 to 8 ml of agitated clotted blood instilled through the guiding needle as it was withdrawn following the biopsy. In the control group (n = 119), the needle was removed without blood instillation. The rate of pneumothorax requiring chest tube placement reduced from 18 to 9% (p = 0.048) with the use of a prophylactic blood patch.

Discussion

Investigation of any therapeutic intervention in pneumothorax is challenging because of the heterogeneity of the population. The expanding number of patients worldwide suffering from a pneumothorax (especially COPD related and iatrogenic) means that persistent air leak will continue to consume an escalating amount of healthcare resource. An effective and minimally invasive procedure (such that patients who are unfit for surgery can benefit) would potentially change the paradigm of care. Clinicians have used ABP for decades, often in anecdotal fashion for patients who had exhausted alternative options. Our literature search demonstrated a lack of high-quality scientific evaluations of the usefulness and safety of the ABP in management of prolonged air leak in pneumothorax. The current data are formed from small retrospective studies or case series, with significant potential for selection and publication bias.

A vicious circle forms when no information exists on the optimal administration of ABP. Studies have followed different protocols with no uniform approach; this heterogeneity of methods has, in turn, reduced the ability to draw firm conclusions about its use. Although ABP has the potential to treat all types of pneumothorax, the vastly different patient groups studied in the existing literature precludes combining the results. Anecdotal successes need to be interpreted with care as air leak often stops spontaneously in patients—a direct causal benefit from ABP from uncontrolled studies must be read with great caution.

Recurrence rates following ABP ranged from 0 to 14%. Most studies reported no recurrence of pneumothorax after a follow-up period ranging from 3 to 48 months. Only one study [35] involving patients with interstitial lung disease (ILD) had a high recurrence rate of 50% following ABP pleurodesis although this was similar to the rate of recurrence in the group treated with chemical pleurodesis (45.5%).

Fig. 1
figure 1

PRISMA flow chart of the included studies

Full size image

The advantage of ABP is that it is readily available all over the world, cheap and well tolerated [29]. An additional advantage is that the lung does not need to be fully inflated before blood instillation (Fig. 2.) [28]. It is relatively safe with empyema being the only major complication reported (0–18%) in a systemic review [22]. One isolated case report of a complicating tension pneumothorax has also been noted [49].

Fig. 2
figure 2

74-year-old gentleman, with prolonged air leak (after 4 days) post chest drain insertion for symptomatic malignant (mesothelioma) pleural effusion (1). He received three consecutive 100 ml doses of autologous blood. The air leak slowed and eventually resolved (2). He remains well at 4 months with no evidence of recurrence (3)

Full size image

Despite these limitations, reports continue to accumulate on the usefulness of ABP. A systematic approach to evaluate its role in the management of ongoing air leak, at least in the common types of pneumothorax, is urgently needed.