Full Terms & Conditions of access and use can be found athttps://www.tandfonline.com/action/journalInformation?journalCode=ypch20

Paediatrics and International Child Health

ISSN: 2046-9047 (Print) 2046-9055 (Online) Journal homepage: https://www.tandfonline.com/loi/ypch20

Bubble CPAP devices for infants and children inresource-limited settings: review of the literature

Alice Won, Daniela Suarez-Rebling, Arianne L. Baker, Thomas F. Burke &Brett D. Nelson

To cite this article: Alice Won, Daniela Suarez-Rebling, Arianne L. Baker, Thomas F. Burke& Brett D. Nelson (2019) Bubble CPAP devices for infants and children in resource-limitedsettings: review of the literature, Paediatrics and International Child Health, 39:3, 168-176, DOI:10.1080/20469047.2018.1534389

To link to this article: https://doi.org/10.1080/20469047.2018.1534389

Published online: 30 Oct 2018.

Submit your article to this journal

Article views: 110

View Crossmark data

Citing articles: 1 View citing articles

https://www.tandfonline.com/action/journalInformation?journalCode=ypch20https://www.tandfonline.com/loi/ypch20https://www.tandfonline.com/action/showCitFormats?doi=10.1080/20469047.2018.1534389https://doi.org/10.1080/20469047.2018.1534389https://www.tandfonline.com/action/authorSubmission?journalCode=ypch20&show=instructionshttps://www.tandfonline.com/action/authorSubmission?journalCode=ypch20&show=instructionshttp://crossmark.crossref.org/dialog/?doi=10.1080/20469047.2018.1534389&domain=pdf&date_stamp=2018-10-30http://crossmark.crossref.org/dialog/?doi=10.1080/20469047.2018.1534389&domain=pdf&date_stamp=2018-10-30https://www.tandfonline.com/doi/citedby/10.1080/20469047.2018.1534389#tabModulehttps://www.tandfonline.com/doi/citedby/10.1080/20469047.2018.1534389#tabModule

Bubble CPAP devices for infants and children in resource-limited settings:review of the literatureAlice Wona, Daniela Suarez-Reblinga, Arianne L. Bakerb,c, Thomas F. Burkea,b and Brett D. Nelson b,d

aDivision of Global Health and Human Rights, Department of Emergency Medicine, Massachusetts General Hospital, Boston, MA, USA;bDepartment of Pediatrics, Harvard Medical School, Boston, USA; cDepartment of Emergency Medicine, Harvard Medical School, Boston,MA, USA; dDivision of Global Health, Department of Pediatrics, Massachusetts General Hospital, Boston, USA

ABSTRACTBackground: Early management of respiratory distress is critical to reducing mortality ininfants and children in resource-limited settings. Bubble continuous positive airway pressure(bCPAP) can offer effective and affordable non-invasive respiratory support.Objective: To determine the best physical components of bubble CPAP circuits for respira-tory support of children in low-resource settings. Methods: Using PubMed, CINAHL andLILACS, studies of any design in any language published before June 2017 which examinedthe physical components of bCPAP circuits were identified and reviewed.Results: After screening, the review included 45 articles: 17 clinical trials, 11 literature reviews,10 technical assessments of bCPAP components, three reports of real-world implementationin low-resource settings, three cost analyses and one case report. There is no ideal bCPAPcircuit for all settings and patients, but some choices are generally better than others indesigning a circuit for low-resource settings. Oxygen concentrators are usually the bestsource of oxygen. As yet, there is no affordable and accurate oxygen blender. Nasal prongsare the simplest patient interface to use with the fewest complications but are not thecheapest option. Expiratory limbs should be at least 1 cm in diameter. Home-made pressuregenerators are effective, safe and affordable.Conclusion: This narrative review found many studies which evaluated the real clinicaloutcomes with bCPAP in the target population as well as technical comparison of bCPAPcomponents. However, many studies were not blinded or randomised and there was sig-nificant heterogeneity in design and outcome measures.

Abbreviations: bCPAP, bubble continuous positive airway pressure; CPAP, continuous posi-tive airway pressure; FiO2, fractional oxygen concentration; HFNC, high-flow nasal cannula;HIC, high-income countries; LMIC, low- and middle-income countries; NP, nasopharyngeal;O2, oxygen; PEEP, positive end-expiratory pressure; PICO, Population, Intervention,Comparison and Outcome

ARTICLE HISTORYReceived 30 April 2018Accepted 7 October 2018

KEYWORDSBubble CPAP; respiratorysupport; respiratory illness;newborns; infants; low-resource setting

Introduction

Nearly three million newborns die each year in thefirst month of life, most of them in low- and middle-income countries (LMIC). The main causes are prema-turity, complications during labour and birth andinfections [1,2], all of which can lead to severe respira-tory distress. In LMIC, as many as 20% of infants withsevere respiratory distress die [3].

Continuous positive airway pressure (CPAP) is widelyused in high-income countries (HIC) and has reducedneonatal morbidity andmortality as well as the need formechanical ventilation and surfactant [4–6]. All forms ofCPAP require the patient to exhale against a constantopening pressure, which produces positive end-expira-tory pressure (PEEP). CPAP, therefore, delivers continu-ous positive pressure into the airways that distends thelungs, prevents alveolar and lung collapse, improvesoxygenation and ventilation and reduces respiratory

fatigue [3,6]. CPAP is typically delivered via mechanicalventilators or commercial pressure drivers in high-resource hospitals; however, it can also be deliveredvia high-flow nasal cannula (HFNC) therapy as well asbubble CPAP (bCPAP) [3].

Both HFNC and bCPAP may be useful in resource-limited settings [7]. The amount of distending pres-sure delivered via HFNC varies and is difficult to mea-sure, whereas in bCPAP, the end of the expiratorylimb of the respiratory circuit is immersed to a depthof water in centimetres that indicates the deliveredCPAP pressure [6]. Furthermore, bubbling generatedby exhalation against the column of water producesnoisy pressure oscillations superimposed over pres-sure fluctuations (stochastic resonance effect) whichpromotes further alveolar recruitment aiding oxyge-nation (Figure 1) [8].

bCPAP can be a low-tech, safe and easy-to-usealternative to conventional CPAP, making it an ideal

CONTACT Alice Won hyewonwon@gmail.com; alice.won@tufts.edu; ahwon@partners.org

PAEDIATRICS AND INTERNATIONAL CHILD HEALTH2019, VOL. 39, NO. 3, 168–176https://doi.org/10.1080/20469047.2018.1534389

© 2018 Informa UK Limited, trading as Taylor & Francis Group

http://orcid.org/0000-0002-5049-1798http://www.tandfonline.comhttp://crossmark.crossref.org/dialog/?doi=10.1080/20469047.2018.1534389&domain=pdf

choice in LMIC. Current commercially available bCPAPcan be relatively affordable at 15% of the cost of amechanical ventilator [9]. However, most bCPAP mod-els remain prohibitively expensive for resource-limitedsettings at prices ranging from approximately US$800to US$6000 [10]. Non-commercial, locally improvisedbCPAP devices assembled from materials found inlocal hospital and community settings may cost aslittle as US$3–5 per single unit [11,12]. Such non-commercial devices, however, may have importantdesign limitations such as the absence of a blenderor a pressure regulator [6,13]. Furthermore, theseprices do not include the cost of oxygen (O2).

While the effectiveness and safety of bCPAP arewell documented [14,15], including in randomisedcontrolled trials in resource-limited settings [16–18],no comprehensive literature review has evaluated theindividual components of bCPAP for feasibility andclinical outcomes in LMIC. This review sought toexamine the evidence and determine the best physi-cal components of bCPAP circuits in LMIC.

Methods

A review protocol was employed that searchedPubMed, CINAHL and LILACS databases using thesearch terms outlined in Table 1. The grey literaturewas also searched using the British Library andGoogle advanced search engine using similar terms.The reference lists of all identified articles were alsosearched and reviewed. The search strategy was

based on the Population, Intervention, Comparisonand Outcome (PICO) format (Table 1).

The literature was searched up to June 2017 withno limits applied to year of publication. The greyliterature was searched up to December 2017. Tomaximize the search sensitivity, search terms pertain-ing to population age and specific outcome cate-gories were not used to generate search results.

Studies that examined various components of abCPAP circuit and described how they affected theO2 delivery to infants and children were included forsystematic review. bCPAP circuit components identifiedfor review throughout various sources were O2 source,inspiratory and expiratory limbs, patient interface andpressure generator (e.g. water reservoir). In addition, rele-vant device parameters such as flow andO2 concentration were also examined as a part of circuitcomponents. Studies that examined the mentioned cir-cuit components in conjunction with non-invasive venti-latory support other than bCPAP were also reviewed.

AW and DS independently reviewed the titles,abstracts and full texts using Covidence (Melbourne,Victoria, Australia) as the main citation managing tool.Each conducted a preliminary screening of titles toexclude studies clearly unrelated to the topic. Theremaining abstracts were screened for appropriate-ness and relevance, followed by perusal of the fulltext for further exclusion of irrelevant papers. Duringeach step of the title, abstract and full-text screening,the two reviewers resolved conflicts by holding dis-cussions and reaching consensus in consultation withthe other authors. Studies not relating to infants or

Figure 1. bCPAP circuit example. A bCPAP circuit includes an oxygen source with humidification canister (A), inspiratory (B) andexpiratory (D) limbs, a patient interface (C) and a pressure generator/water reservoir (E). The depth at which the expiratory limbof the bCPAP circuit is placed underwater determines the amount of pressure generated.

PAEDIATRICS AND INTERNATIONAL CHILD HEALTH 169

children were excluded. No studies were excluded onthe basis of language or design.

Various causes of respiratory distress were considered,including respiratory distress syndrome, pneumonia, sep-sis and transient tachypnoea of the newborn, but studiesaddressing congenital anomalies and structural patholo-gies were excluded. Studies exclusively involving HIC andthose examining the efficacy of commercial devices with-out adding to the originally published data regardingtheir individual components were also excluded. A sum-mary of the findings from the included studies was com-piled in Microsoft Excel (Redmond, WA, USA).

Results

After screening, the review included 45 articles: 17clinical trials, 11 literature reviews, 10 technical assess-ments of bCPAP component, three reports of real-world implementation in low-resource settings, threecost analyses and one case report (Figure 2).

Oxygen source

The two common sources of O2 for bCPAP areO2 cylinders and O2 concentrators. O2 cylinders con-tain liquid O2 that is distilled at very low temperaturesand high pressures in a special facility and, therefore,must be transported back and forth from the hospitalfor regular refilling. O2 concentrators are suitcase-sized electrically powered machines which draw inambient air and extract nitrogen, leaving 90–95%pure O2 for use [19]. A simple comparison is sum-marised in Table 2.

Field studies of O2 concentrators implemented inlarge-scale programmes in low-resource settings found

that the majority remained in use years afterwards [20–22]. However, these studies were typically using stan-dard-flow oxygen delivery and not the higher flow ratestypically required in bCPAP. Under the demands ofbCPAP, many concentrators fall short of the requisiterobustness. One study evaluated seven commerciallyavailable concentrators and found that only one per-formed acceptably well in the conditions specified byWHO for low-resource settings, although the study didnot specifically assess performance at the higher flowrates used in bCPAP [13]. In a randomised clinical trial ofbCPAP in Bangladesh, the study’s specific concentratormodel failed 21% of the time during bCPAP, requiringback-up O2 supplies to continue patient treatment [16].

The main challenge with O2 concentrators is theneed for constant electricity [23] and solar power isemerging as a potential solution [24,25]. Anothersolution reported is an O2 reservoir which consists ofa non-elastic balloon connected to an O2 concentratorwhich fills the reservoir and can be used in the eventof power failure [26].

O2 flow meter and flow splitter

Flow rates in bCPAP circuits affect the delivered pressure.Flow can be regulated either by controlling the amountof O2 from an O2 source or at the point of blending airand O2 [27]. The literature review did not find any directcomparison of different flowmeters, although the major-ity of reports of bCPAP included flow meters that coulddeliver child-appropriate flow rates.

Flow through a bCPAP circuit is influenced by thecircuit’s diameter, length and integrity, including thedegree of seal at the nasal interface. The approximateflow can be visually assessed by observing the rate of

Table 1. PICO search strategy used for literature search.a

PICO terms Description Search terms

Population Newborns and infants with respiratory distress in low- and middle-income countries

Childb, childrenb, infant(s)b, infancyb, paediatricb, paediatricb,neonate(s)b, neonatalb, newbornb, developing country,developing countries, underdeveloped country, underdevelopedcountries, low income country, low income countries, middleincome country, middle income countries, resource poor,resource limited, low resource

Intervention Bubble CPAP Bubble CPAP, bubbling CPAP, bCPAP, bubble continuous positiveairway pressure

Comparison Other forms of non-invasive oxygen therapy, including nasal CPAPand standard oxygen via nasal cannula

Continuous positive airway pressure, positive-pressure respiration,positive airway pressure device, nCPAP, nasal continuouspositive airway pressure, oxygen therapy, respiratory supportdevice, nasal cannula, nasal cannulae, non-invasive respiratorysupport, non-invasive ventilation

Outcome Treatment failure, treatment complications, improvement inrespiratory parameters (e.g. respiratory rate), rate of intubation,rate of invasive/mechanical ventilation, severity of respiratorydistress, mortality/survival to discharge

Outcome-based search termsb were not applied in the searchprocess

Question For newborns and infants with respiratory distress treated inresource-limited settings, what are the best components of abubble CPAP circuit and other related non-invasive ventilationmethods that contribute to improved outcomes?

aSimilar search terms were used for other databases with adaptations as needed. bAge-specific search terms and outcome-based search terms were notapplied in the search process to maximize the search sensitivity.

170 A. WON ET AL.

bubbling in the circuit. While strategies for choosingflow rates were beyond the scope of this review, onestudy reported that using a fixed flow delivered accu-rate pressures to patients as opposed to titrating flowto produce bubbling [27].

Managing O2 flow is also important for conservingO2 supplies [28]. If multiple children can tolerate thesame flow rate, a flow splitter device can be used tomaximize O2 supplies and decrease the overall cost of

O2 [22,29,30]. CPAP usually requires O2 flow rates of2–10 L/min, whereas a concentrator can usually pro-vide only up to 5–8 L/min and, when split, will have aproportionally lower flow. For example, one groupmeasured actual flow rates for each identical limb ofa four-way flow-splitter and found that a concentratorset at 4 L/min would deliver 0.5 L/min to each limb[22]. One innovative system used an O2 concentratorfor CPAP by adding an air compressor via a Y-piece to

Table 2. Comparison of oxygen delivery systems.Requirements Cylinders Concentrator References

Infrastructure Reliable transportation to/from central processing centres Continuous electrical supply 20,22,23,30,50Additionalequipment

Pressure regulator (~US$200/cylinder), flow meter (~US$400), humidifier

None 20,22

Cost ~US$1500 for 1 million L oxygen; cost of cylinder may behigher in LMICOngoing costs of transportation for frequent refilling;standard cylinders lasting 2–3 days with continuous useA 2-year operational cost of ~US$168,500a (largeprovincial hospital) and ~US$102,000a (small districthospitalb) in Papua New Guinea~2.5–15× more expensive overall

~US$1500 per machine which can produce ~1 million Loxygen in 6 mthsModerate up-front costs of procuring initial equipmentand installation, but small ongoing costs (electricity,maintenance)A 2-year operational cost of ~US$68,300a (largeprovincial hospital) and ~US$10,100a (small districthospitalb) in Papua New Guinea

20,22,23,50

Maintenance Minimal; in central facility where oxygen is compressed athigh pressure and low temperature

Local maintenance and repair, spare parts 20,21,22

aApproximated by conversion from values in Papua New Guinea’s currency, Kina (PGK).bOperational cost in a small district hospital not including the cost of an anaesthetic machine required for major surgery.

Figure 2. Flow chart of literature search results.

PAEDIATRICS AND INTERNATIONAL CHILD HEALTH 171

increase the total flow rate, making it suitable forCPAP [29]. No studies comparing use of a flow splitterwith not using a flow splitter were identified.

O2 blender

In low-resource settings, air–O2 blenders that incorpo-rate ambient room air and pure O2 to supply preciselymixed FiO2 are not usually available [11,12,28,29,31,32].Without an O2 blender, several authors reported noability to regulate FiO2 or difficulty in maintainingtarget FiO2 as well as differences between predictedand measured FiO2 [11,28,31–33]. Some commerciallyavailable bCPAP systems do include an air–O2 blender,which allows for use with premature newborns, buthomemade bCPAP usually does not [6]. Air–O2 blenders also require a high pressure O2 sourcebeyond the capability of most O2 concentrators,which produce low-pressure O2 [33].

For patients such as premature newborns who areat particularly high risk of O2 toxicity and retinopathyof prematurity, WHO recommends blended O2 withFiO2 of 0.3 or room air (0.21 FiO2) [34,35]. However,data show that excessive use of oxygen in adults isalso associated with excess mortality [12,36].

Several authors reported innovative systems toadjust the concentration of O2 without a blender. AY-tubing set-up delivers a mixture of air and O2 fromtwo separate supplies with independent flow metersand can be used to change the relative concentrationof O2 [29,33]. Kaur et al. tested calculatedO2 concentrations of 19–95% using such a Y-tubingsystem and found actual O2 concentrations of 21–98%with variable accuracy [33]. Alternatively, an air pumpcan add O2 from a concentrator to air, with separateflow control for pure O2 to control the FiO2 [27,37].

Another option is an entrainment device whichuses a small jet of O2 and draws in ambient airthrough an adjustable inlet hole, thus mixing theO2 and air to achieve a set concentration. However,one common entrainment device designed for adultsdid not deliver an appropriate O2 concentration orflow when used with nasal cannula-like tubing at achild-appropriate flow rate [28].

Patient interface

Four patient interface devices for bCPAP were identi-fied: nasal prongs/cannulas, nasal catheters, nasophar-yngeal (NP) catheters and nasal/face mask [3,7]. Acomparison of their basic characteristics is shown inTable 3. In this review, the most common patientinterface used for bCPAP was nasal prongs. Nasaland NP catheters may also be used for CPAP,although they require more nursing intervention andare more prone to complications [33,38,39].

When used to deliver low-flow O2, nasal cannulason average require higher O2 flow rates than NPcatheters to achieve the same partial pressure ofO2 [12]. At equal flow rates, NP catheters deliver thehighest FiO2 compared with nasal catheters andprongs [38].

NP catheters produce increased PEEP with lowerO2 flow rates than nasal prongs [40]. Larger NP cathe-ters were found to produce PEEP in infants; for exam-ple, size 8 Fr catheters produce PEEP (6.3 cm at 1 L/min, 10.6 at 2 L/min) while size 6 Fr do not [12,40].

Nasal prongs are the easiest to use with the leastserious complications that can occur with catheterssuch as displacement into the airway and gastric dis-tension [38,41]. High-flow air through nasal prongsdoes require a humidification device, however, as do

Table 3. Comparison of patient interface devices for bCPAP.Nasal prong/

cannula Nasal catheter Nasopharyngeal catheter Face mask References

O2 flowrequired*

1.26× ~1× 1× Higher flow requirements 12,41

Cost per device ~US$2–5 ~US$0.10 ~US$0.10 Higher cost 38Humidificationrequired

Yes Yes Yes No 12,51

Risk of airwayobstructionby mucus

Low/Slight High High Low 38,41

Complications Dislodgement;tube breakage

Nasal bleeding; small risk ofdisplacement and gastricdistension

Displacement and gastricdistention; airwayperforation

Carbon dioxide accumulation 12,38

Limitations Not easy todetermineprecise FiO2

Higher nursing demand Highest nursing demand;highest complicationrate

Special ordering required – notgenerally available; incompatiblewith feeding tube

38,40,41

Other Reservoir prongscan conserveoxygen

Nasogastric tube required Nasogastric tuberequired

45

*O2 flow required compared with nasopharyngeal catheter.

172 A. WON ET AL.

nasal and NP catheters since they bypass the nasalpassages which typically provide humidification.Studies of bCPAP have included humidified O2 foruse with all three interfaces [3,6,33].

Several comparisons of different types of nasalprongs were identified. Short binasal prongs aremore effective in CPAP than single prongs, with lessresistance to flow, and they are more effective atpreventing re-intubation [7,42,43]. Commercially avail-able prongs are equally effective for reaching targetPEEP and O2 saturation [43]. When tested on a lungmodel, standard infant- or newborn-sized nasalprongs delivered consistently higher mean CPAPthan thinner-walled prongs [44].

Nasal prongs are 20–50 times more expensive thancatheters at an estimated cost of US$2–5 per set ofprongs compared with US$0.10 for a paediatric naso-gastric tube or suction tube which can be used as acatheter [38]. If O2 conservation is a priority, reservoircannulas include an inline compliant reservoir thatcaptures exhaled O2 between breaths and delivers itduring inhalation, which can further conserveO2 supplies [45].

Expiratory limb

The expiratory limb for bCPAP is a tube of non-col-lapsible plastic leading from the patient interface tothe pressure generator where it is immersed in water.

When different diameters of the expiratory limbwere compared in a lung model, a larger expiratorylimb (>10 mm) with greater depth caused greateroscillations in pressure and volume, especially ininfants with low lung compliance, which shouldimprove gas exchange [46].

The expiratory limb should be emptied at least every2–3 h since condensation collecting in the limbincreased the delivered pressure beyond set levels [47].

Water reservoir or pressure generator

The water reservoir used to generate pressure may behome-made or commercially produced. Home-madewater reservoirs consist of a water-filled bottle (e.g.intravenous solution bottle, shampoo bottle or glassgraduated cylinder). The expiratory limb is submergedand stabilised in the bottle, and the water columnheight within the limb determines the pressure.Several field studies have reported the feasibility,affordability and effectiveness of home-made pres-sure generators in LMIC [6,11,16,27,29,32,48].

Commercial models have different mechanisms forcontrolling the pressure generated by the water col-umn [49]. Fisher and Paykel’s system has a rigid tubemoulded to fit inside the reservoir which generatespressure. Babi Plus has a rotation mechanism insidethe expiratory limb that changes the depth and

controls pressure. WaterPAP uses a water bottle witha corrugated tube with a plastic lid that holds thecorrugated tubing in place.

Home-made water reservoirs tested in low-resourcesettings accurately delivered pressures within 1 cmacross a range of pre-set pressures [32]. In one studycomparing home-made systems with commercial ones,a home-made pressure generator had more variableairway pressure and volume oscillations with the leastincrease in pressure at increasing flows but the home-made system is not as foolproof [49].

One lung model using different home-made pres-sure generators found that a smaller bottle such as a500-ml graduated cylinder should be used to increasepressure and volume oscillations [46].

Discussion

The ideal bCPAP device for low-resource settingswould be safe, effective, affordable, reusable, readilyavailable and simple to use. On reviewing the indivi-dual components of bCPAP in LMIC, the evidencepoints to several superior choices while highlightingareas for further studies with potential for innovationand development.

O2 concentrators are safe and cost-efficient, poten-tially making them the best choice as a source of O2 inmost low-resource settings. They do not require layersof widescale infrastructure, whereas for O2 cylinders,there is a chain of infrastructure from the financing ofhigh-energy production of liquid oxygen at an oxygenplant to reliable road and transportation systems[20,30,50]. The need for an uninterrupted power sup-ply, however, can be a limiting factor in LMIC. Solarpower or O2 reservoirs may allow concentrator useeven without reliable electricity. Furthermore, whilethe initial cost of installing a solar-powered system ishigh, operational costs tend to be low with only airand sun being required [24]. It is essential, however,to ensure the purchase and implementation ofO2 concentrators that meet the performance stan-dards outlined by the WHO guideline for technicalspecifications for O2 concentrators in low-resourcesettings [34]. It is also essential to thoroughly assessa facility’s O2 requirements, conduct local training inmaintenance and repair and ensure the presence ofback-up power or O2 supply.

Traditional commercially available O2 blenders aretoo expensive for low-resource settings and home-made blending systems have limited precision andaccuracy. Low-resource settings need an affordableand reliable O2 blender in order to expand the safeuse of bCPAP, especially for premature infants whoare vulnerable to retinopathy of prematurity owing toO2 toxicity. O2 concentrators with built-in blenders arecurrently the best choice for premature infants.

PAEDIATRICS AND INTERNATIONAL CHILD HEALTH 173

However, an air entrainment device designed on thebasis of the Venturi effect – a pressure differentialgenerated by a simple manipulation of oxygen inletand outlet orifices that can lead to variable air–oxy-gen blending capacities – is a potentially powerfuland revolutionary tool [28]. Future studies shouldoptimise the design and application of such entrain-ment devices with attention to cost-efficiency as wellas reported limitations.

For the patient interface, nasal prongs are the sim-plest to use with the fewest serious complications.The most common complications of nasal prongsare dislodgement and nasal irritation, whereas com-plications associated with catheters include displace-ment into the lower airways with the risk of airwayperforation and gastric distension. However, nasalprongs do require more O2 flow, and they are muchmore expensive than nasal and NP catheters. Anotherpossible limitation with narrow, high-resistance nasalprongs is a pseudo-CPAP bubbling effect. Our reviewrevealed a lack of data that specifically analyses thebubbling effect of bCPAP with variables that deter-mine the resistance of the circuit.

Home-made pressure generators are reliable, accu-rate and affordable and in some respects (such as increating oscillation in pressure and volume whichencourages gas exchange) may be superior to com-mercially available devices. The diameter should benarrow, approximately 5–6 cm, to increase pressureand volume oscillation. Lastly, the expiratory limbshould be at least 1 cm in diameter for increasedpressure and volume oscillations to facilitate gasexchange in the lung.

The limitations of this review include its limitedscope. Best practices for system-wide implementationof bCPAP and training of clinical staff were notdirectly part of the search strategy, although theauthors discussed the importance of good trainingin its use as well as maintenance and repair. Oneelement of safe bCPAP implementation, as with anyuse of supplemental O2, is monitoring with pulseoximetry. WHO guidelines support this recommenda-tion [34,51,52]. This review also included a limitedgrey literature search and there is probably a wealthof information from innovative providers not widelypublished and not seen in this review. Finally, this wasa narrative review that aimed to assess all of therelevant data published but it did not include sys-tematic evaluation of the quality of the literatureincluded.

CPAP is a safe and effective method of treatinginfants and young children with life-threateningrespiratory distress and can be successfully adaptedfor use in resource-limited settings. This literature

review found that the most feasible, safe and cost-effective bCPAP system in LMIC would be anO2 concentrator connected to short binasal prongswith an attached humidifier, with a large-diameterexpiratory limb submerged into a small-diameterhome-made pressure generator such as a graduatedcylinder. There is a need for an affordable O2-blending device to expand the use of bCPAP to pre-mature infants and others. Successful bCPAP imple-mentation will depend on the establishment ofreliable infrastructure such as electricity, strong clin-ical training and local training in device maintenanceand repair.

Disclosure statement

No potential conflict of interest was reported by theauthors.

Notes on contributors

Alice Won is a fourth-year medical student at TuftsUniversity School of Medicine and spent a research concen-tration year at Massachusetts General Hospital's Division ofGlobal Health and Human Rights. She is interested in opera-tional research utilizing existing healthcare resources andsystems to address gaps in access to quality care in com-munities at home and abroad.

Daniela Suarez-Rebling is a program coordinator at theMassachusetts General Hospital Division of Global Healthand Human Rights. She is interested in public healthresearch, specifically identifying how to scale innovationsto impact vulnerable populations and to address healthdisparities in populations.

Arianne L. Baker is a resident in Pediatrics at theMassachusetts General Hospital and Harvard MedicalSchool. Her research work focuses on leadership develop-ment and medical education. She has also worked on qual-ity improvement for mothers and their infants in a Boston-area pediatric primary care community health center. Herclinical interests include pediatric acute care.

Thomas F. Burke is Chief of the Division of Global Healthand Human Rights in the Department of EmergencyMedicine at the Massachusetts General Hospital. He is asenior emergency physician and an Associate Professor atHarvard Medical School. He has extensive experience instrategic planning and development of service and researchcapacity for global health interventions, most notably forcommunity-based maternal and infant health delivery. Hehas been a leader in the nonprofit sector directing overseashealth interventions since 1994.

Brett D. Nelson is Associate Professor and global healthpediatrician at Massachusetts General Hospital andHarvard Medical School. He has been involved in clinicalcare and program management in dozens of low- and mid-dle-income countries. His research interests are healthcareprovision, development, research, and advocacy for

174 A. WON ET AL.

vulnerable populations, particularly newborns and childrenin settings affected by poverty, conflict, or disaster. Hedirects Harvard Medical School’s course on global healthand serves as editor of a textbook on clinical global health.

ORCID

Brett D. Nelson http://orcid.org/0000-0002-5049-1798

References

[1] UNICEF, WHO, World Bank Group and United Nations.Levels and trends in child mortality report 2017.Accessed October 18, 2018. https://www.unicef.org/publications/index_101071.html. Updated 2017.

[2] Lawn JE, Blencowe H, Oza S, et al. Every newborn:progress, priorities, and potential beyond survival.Lancet. 2014;384:189–205.

[3] Duke T. CPAP: a guide for clinicians in developingcountries. Paediatr Int Child Health. 2014;34:3–11.

[4] Jensen EA, Chaudhary A, Bhutta ZA, et al. Non-inva-sive respiratory support for infants in low- and mid-dle-income countries. Semin Fetal Neonatal Med.2016;21:181–188.

[5] Ho JJ, Subramaniam P, Davis PG. Continuous distend-ing pressure for respiratory distress in preterminfants. Cochrane Database Syst Rev. 2015;7:CD002271.

[6] Dewez JE, van den Broek N. Continuous positive air-way pressure (CPAP) to treat respiratory distress innewborns in low- and middle-income countries. TropDoct. 2017;47:19–22.

[7] Gupta N, Saini SS, Murki S, et al. Continuous positiveairway pressure in preterm neonates: an update ofcurrent evidence and implications for developingcountries. Indian Pediatr. 2015;52:319–328.

[8] Pillow JJ, Hillman N, Moss TJ, et al. Bubble continuouspositive airway pressure enhances lung volume andgas exchange in preterm lambs. Am J Respir Crit CareMed. 2007;176:63–69.

[9] Wyatt J. Appropriate medical technology for perinatalcare in low-resource settings. Ann Trop Paediatr.2008;28:243–251.

[10] PATH, USAID. Continuous positive airway pressure(CPAP) – Guide to selection, 2014 Accessed October18, 2018. https://www.path.org/publications/files/DT_guide_to_select_cpap.pdf

[11] Daga S, Mhatre S, Borhade A, et al. Home-made con-tinuous positive airways pressure device may reducemortality in neonates with respiratory distress in low-resource setting. J Trop Pediatr. 2014;60:343–347.

[12] Frey B, Shann F. Oxygen administration in infants.Arch Dis Child Fetal Neonatal Ed. 2003;88:84.

[13] Peel D, Neighbour R, Eltringham RJ. Evaluation ofoxygen concentrators for use in countries with lim-ited resources. Anaesthesia. 2013;68:706–712.

[14] Martin S, Duke T, Davis P. Efficacy and safety ofbubble CPAP in neonatal care in low and middleincome countries: a systematic review. Arch DisChild Fetal Neonatal Ed. 2014;99:F495–F504.

[15] Thukral A, Sankar MJ, Chandrasekaran A, et al. Efficacyand safety of CPAP in low- and middle-income coun-tries. J Perinatol. 2016;36(Suppl 1):S21–S28.

[16] Chisti MJ, Salam MA, Smith JH, et al. Bubble continuouspositive airway pressure for children with severe

pneumonia and hypoxaemia in Bangladesh: an open,randomised controlled trial. Lancet. 2015;386:1057–1065.

[17] Tagare A, Kadam S, Vaidya U, et al. Bubble CPAPversus ventilator CPAP in preterm neonates withearly onset respiratory distress: a randomized con-trolled trial. J Trop Pediatr. 2013;59:113–119.

[18] Yagui AC, Vale LA, Haddad LB, et al. Bubble CPAPversus CPAP with variable flow in newborns withrespiratory distress: a randomized controlled trial. JPediatr (Rio J). 2011;87:499–504.

[19] Charburn RL, Williams TJ. Performance comparison of4 portable oxygen concentrators. Respiratory Care.2010;55:433–442.

[20] Dobson MB. Oxygen concentrators and cylinders. Int JTuberc Lung Dis. 2001;5:520–523.

[21] La Vincente SF, Peel D, Carai S, et al. The functioningof oxygen concentrators in resource-limited settings:a situation assessment in two countries. Int J TubercLung Dis. 2011;15:693–699.

[22] Matai S, Peel D, Wandi F, et al. Implementing anoxygen programme in hospitals in Papua NewGuinea. Ann Trop Paediatr. 2008;28:71–78.

[23] Mokuolu OA, Ajayi OA. Use of an oxygen concentratorin a Nigerian neonatal unit: economic implicationsand reliability. Ann Trop Paediatr. 2002;22:209–212.

[24] Turnbull H, Conroy A, Opoka RO, et al. Solar-poweredoxygen delivery: proof of concept. Int J Tuberc LungDis. 2016;20:696–703.

[25] Nyende S, Conroy A, Opoka RO, et al. Solar-poweredoxygen delivery: study protocol for a randomizedcontrolled trial. Trials. 2015;16:297.

[26] Rassool RP, Sobott BA, Peake DJ, et al. A low-pressureoxygen storage system for oxygen supply in low-resource settings. Respir Care. 2017;62:1582–1587.

[27] Brown J, Machen H, Kawaza K, et al. A high-value, low-cost bubble continuous positive airway pressure systemfor low-resource settings: technical assessment and initialcase reports. PLoS One. 2013;8:e53622.

[28] Coghill M, Ambalavanan N, Chatburn RL, et al.Accuracy of a novel system for oxygen delivery tosmall children. Pediatrics. 2011;128:382.

[29] van den Heuvel M, Blencowe H, Mittermayer K, et al.Introduction of bubble CPAP in a teaching hospital inMalawi. Ann Trop Paediatr. 2011;31:59–65.

[30] Howie SRC, Hill S, Ebonyi A, et al. Meeting oxygenneeds in Africa: an options analysis from the Gambia.Bull WHO. 2009;87:763–771.

[31] Jayashree M, KiranBabu H, Singhi S, et al. Use of nasalbubble CPAP in children with hypoxemic clinicalpneumonia – Report from a resource limited set-up.J Trop Pediatr. 2016;62:69–74.

[32] McAdams RM, Hedstrom AB, DiBlasi RM, et al.Implementation of bubble CPAP in a rural Ugandanneonatal ICU. Respir Care. 2015;60:437–445.

[33] Kaur C, Sema A, Beri RS, et al. A simple circuit to deliverbubbling CPAP. Indian Pediatr. 2008;45:312–314.

[34] World Health Organization. Technical specificationsfor oxygen concentrators. Geneva: WHO; 2015.Accessed October 18, 2018. http://apps.who.int/iris/bitstream/handle/10665/199326/9789241509886_eng.pdf?sequence=1

[35] World Health Organization. WHO recommendationson interventions to improve preterm birth outcomes.Geneva: WHO; 2015. Accessed October 18, 2018.http://apps.who.int/iris/bitstream/handle/10665/183037/9789241508988_eng.pdf?sequence=1

PAEDIATRICS AND INTERNATIONAL CHILD HEALTH 175

https://www.unicef.org/publications/index_101071.htmlhttps://www.unicef.org/publications/index_101071.htmlhttps://www.path.org/publications/files/DT_guide_to_select_cpap.pdfhttps://www.path.org/publications/files/DT_guide_to_select_cpap.pdfhttp://apps.who.int/iris/bitstream/handle/10665/199326/9789241509886_eng.pdf?sequence=1http://apps.who.int/iris/bitstream/handle/10665/199326/9789241509886_eng.pdf?sequence=1http://apps.who.int/iris/bitstream/handle/10665/199326/9789241509886_eng.pdf?sequence=1http://apps.who.int/iris/bitstream/handle/10665/183037/9789241508988_eng.pdf?sequence=1http://apps.who.int/iris/bitstream/handle/10665/183037/9789241508988_eng.pdf?sequence=1

[36] Chu DK, Kim LH, Young PJ, et al. Mortality and mor-bidity in acutely ill adults treated with liberal versusconservative oxygen therapy (IOTA): a systematicreview and meta-analysis. Lancet. 2018;391:1693–1705.

[37] Kawaza K, Machen HE, Brown J, et al. Efficacy of a low-cost bubble CPAP system in treatment of respiratorydistress in a neonatal ward in Malawi. PLoS One.2014;9:e86327.

[38] Muhe L, Webert M. Oxygen delivery to children withhypoxaemia in small hospitals in developing coun-tries. Int J Tuberc Lung Dis. 2001;5:527–532.

[39] Walk J, Dinga P, Banda C, et al. Non-invasive ventila-tion with bubble CPAP is feasible and improvesrespiratory physiology in hospitalised Malawian chil-dren with acute respiratory failure. Paediatr Int ChildHealth. 2016;36:28–33.

[40] Frey B, McQuillan PJ, Shann F, et al. Nasopharyngealoxygen therapy produces positive end-expiratorypressure in infants. Eur J Pediatr. 2001;160:556–560.

[41] Weber MW, Palmer A, Oparaugo A, et al. Comparisonof nasal prongs and nasopharyngeal catheter for thedelivery of oxygen in children with hypoxemiabecause of a lower respiratory tract infection. JPediatr. 1995;127:378–383.

[42] De Paoli AG, Davis PG, Faber B, et al. Devices andpressure sources for administration of nasal con-tinuous positive airway pressure (NCPAP) in pre-term neonates. Cochrane Database Syst Rev.2008;1:CD002977.

[43] Gupta S, Donn SM. Continuous positive airway pres-sure: physiology and comparison of devices. SeminFetal Neonatal Med. 2016;21:204–211.

[44] Bailes SA, Firestone KS, Dunn DK, et al. Evaluating theeffect of flow and interface type on pressures

delivered with bubble CPAP in a simulated model.Respir Care. 2016;61:333–339.

[45] Wu G, Wollen A, DiBlasi RM, et al. Reservoir cannulasfor pediatric oxygen therapy: a proof-of-conceptstudy. Int J Pediatr. 2016;2016:9214389.

[46] Wu C, Lee C, Yuh Y, et al. Influence of changing thediameter of the bubble generator bottle and expira-tory limb on bubble CPAP: an in vitro study. PediatrNeonatol. 2012;53:359–365.

[47] Youngquist TM, Richardson CP, Diblasi RM. Effects ofcondensate in the exhalation limb of neonatal circuitson airway pressure during bubble CPAP. Respir Care.2013;58:1840–1846.

[48] Myhre J, Immaculate M, Okeyo B, et al. Effect oftreatment of premature infants with respiratory dis-tress using low-cost bubble CPAP in a rural Africanhospital. J Trop Pediatr. 2016;62:385–389.

[49] Poli JA, Richardson CP, DiBlasi RM. Volume oscillationsdelivered to a lung model using 4 different bubbleCPAP systems. Respir Care. 2015;60:371–381.

[50] Duke T, Peel D, Wandi F, et al. Oxygen supplies forhospitals in Papua New Guinea: a comparison of thefeasibility and cost-effectiveness of methods for dif-ferent settings. PNG Med J. 2010;53:126–138.

[51] World Health Organization. Oxygen therapy for children:a manual for health workers. Geneva: WHO; 2016.Accessed October 18, 2018. http://apps.who.int/iris/bitstream/handle/10665/204584/9789241549554_eng.pdf;jsessionid=33AA54C2D67C3154A1B85DCE3DDABA32?sequence=1

[52] World Health Organization. Pulse oximetry trainingmanual. Geneva: WHO; 2011. Accessed October 18,2018. http://www.who.int/patientsafety/safesurgery/pulse_oximetry/who_ps_pulse_oxymetry_training_manual_en.pdf

176 A. WON ET AL.

http://apps.who.int/iris/bitstream/handle/10665/204584/9789241549554_eng.pdf;jsessionid=33AA54C2D67C3154A1B85DCE3DDABA32?sequence=1http://apps.who.int/iris/bitstream/handle/10665/204584/9789241549554_eng.pdf;jsessionid=33AA54C2D67C3154A1B85DCE3DDABA32?sequence=1http://apps.who.int/iris/bitstream/handle/10665/204584/9789241549554_eng.pdf;jsessionid=33AA54C2D67C3154A1B85DCE3DDABA32?sequence=1http://apps.who.int/iris/bitstream/handle/10665/204584/9789241549554_eng.pdf;jsessionid=33AA54C2D67C3154A1B85DCE3DDABA32?sequence=1http://www.who.int/patientsafety/safesurgery/pulse_oximetry/who_ps_pulse_oxymetry_training_manual_en.pdfhttp://www.who.int/patientsafety/safesurgery/pulse_oximetry/who_ps_pulse_oxymetry_training_manual_en.pdfhttp://www.who.int/patientsafety/safesurgery/pulse_oximetry/who_ps_pulse_oxymetry_training_manual_en.pdf

AbstractIntroductionMethodsResultsOxygen sourceO2 flow meter and flow splitterO2 blenderPatient interfaceExpiratory limbWater reservoir or pressure generator

DiscussionDisclosure statementNotes on contributorsORCIDReferences