All trials using random or quasi-random patient allocation, in which preoxygenation prior to tracheal suctioning was compared to suctioning without preoxygenation.

Newborn infants receiving ventilatory support via an endotracheal tube who received endotracheal suctioning.

Preoxygenation immediately prior to endotracheal suctioning.

Primary outcomes

Secondary outcomes

Searches included the Oxford Database of Perinatal Trials, Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 2, 2007), MEDLINE (1966 - September 2007), CINAHL (1982 - September 2007) using MeSH term infant-newborn and text terms oxygen*, suction*, preoxygenation, pre-oxygenation and premature.

In January of 2009 we updated the search as follows:
MEDLINE (search via PubMed), CINAHL, EMBASE and The Cochrane Library were searched from 2007 to January 31, 2009.
Search terms: oxygen*, suction*, preoxygenation, pre-oxygenation. Limits: human, newborn infant and clinical trial.
No language restrictions were applied.

The search included previous reviews including cross references, abstracts in conferences and symposia proceedings, and contact with expert informants.

Clinical trials registries were also searched for ongoing or recently completed trials (clinicaltrials.gov; controlled-trials.com; and who.int/ictrp)

All randomized and quasi-randomized controlled trials fulfilling the selection criteria described in the previous section were included. The investigators reviewed the results of the search and separately selected the studies for inclusion. The review authors resolved any disagreement by discussion.

Two review authors (MP, VF) independently extracted data, then compared and resolved differences. For each study, final data was entered into RevMan by one review author and then checked by a second review author.

Four major sources of potential bias and methods of avoidance of these biases were considered when assessing trial quality; 1) Selection bias - blinding of randomization; 2) Performance bias - blinding of intervention; 3) Attrition bias - complete follow-up; 4) Detection bias - blinding of outcome assessment. The Cochrane Neonatal Group bases its quality assessments on systematic assessment of the opportunity for each of these biases to arise. Thus, the review authors judged from the report of the trial whether each of the criteria (methods of avoidance) was met. Each criterion was given a rating of either A if Yes (Adequate), B if Can't Tell (Unclear), or C if No (Inadequate). The quality assessment rating included in the Table of Included Studies refers to blinding of randomization alone.

The methodological quality of each trial was reviewed by two review authors (MP, VF) who then compared and resolved differences. Additional information was sought for all identified potentially eligible studies on randomization method. However, at the time of the review no correspondence was received.

For the update in 2009, the risk of bias table was completed in order to address the following questions:

1. Sequence generation: Was the allocation sequence adequately generated? 
2. Allocation concealment: Was allocation adequately concealed? 
3. Blinding of participants, personnel and outcome assessors: Was knowledge of the allocated intervention adequately prevented during the study? At study entry? At the time of outcome assessment? 
4. Incomplete outcome data: Were incomplete outcome data adequately addressed? 
5. Selective outcome reporting: Are reports of the study free of suggestion of selective outcome reporting? 
6. Other sources of bias: Was the study apparently free of other problems that could put it at a high risk of bias?

Relative risk (RR) was computed for categorical data and Mean difference (MD) for data measured on a continuous scale, with 95% confidence intervals (CI) presented for all reported outcomes.

If multiple trials were identified, we planned to estimate the treatment effects of individual trials and examine heterogeneity between trials by inspecting the forest plots and quantifying the impact of heterogeneity using the I-squared statistic. If we detected statistical heterogeneity, we planned to explore the possible causes (for example, differences in study quality, participants, intervention regimens, or outcome assessments) using post hoc sub group analyses.

If multiple studies were identified and meta-analysis was judged to be appropriate, the analysis would have been performed using Review Manager software (RevMan 5, Cochrane Collaboration). For estimates of typical relative risk and risk difference, we planned to use the Mantel-Haenszel method. For measured quantities, we planned to use the inverse variance method. All meta-analyses were to be done using the fixed effect model. However, due to insufficient data, meta-analysis was not able to be conducted.

Subgroup analyses are planned to determine whether the results differ by:

Population (newborn infants) :

• by gestational age < 30 weeks, < 34 weeks and < 37 weeks; and
• by disease, infants with chronic lung disease compared to those without.

Intervention (different techniques of endotracheal suctioning):

• with or without disconnection from the ventilator;
• increased mechanical ventilation; and
• use of manual ventilation and chest wall vibrations or percussion.

Four studies were identified as potentially eligible for inclusion in this review. Two studies were excluded (Cabal 1979 and Graff 1987) and one study is awaiting further classification (Gonzalez 2005). Therefore, this review includes one trial (Walsh 1987).

Walsh 1987: The included study utilised a cross-over design and involved 21 infants. Participants were preterm infants with gestational age 31.4 (± 2.5) weeks, birthweight 1586 (± 682) grams and postnatal age at enrolment 4 (± 4.5) days [(Mean (SD)]. All infants were mechanically ventilated for respiratory distress with FiO2 > 0.30.

In this study, two suctioning procedures were compared. Procedure A included the following sequence of events: chest vibration using a hand held vibrator for two minutes, instillation of normal saline, reconnection to ventilator for 30 seconds, endotracheal tube suctioning with infant's head to one side, returned to ventilator for 30 seconds, second endotracheal suctioning with head on the opposite side, and then returned to ventilator. Procedure B included the same sequence of events as described in procedure A with the addition of preoxygenation, which involved an increase in FiO2 until Tc PO2 reached between 90 -100 mmHg and stabilised for two minutes. Following suction the FiO2 was gradually reduced to base line setting. Each infant was studied for one suction in each of the two procedures.

The trial measured effects of preoxygenation prior to suctioning on hypoxemia and bradycardia. Transcutaneous oxygen (Tc PO2) values were reported at ten time points over a suctioning procedure. In this review, data is presented for three of these measurement points, chosen because of clinical relevance. These points are: 1) end of first endotracheal suction 2) end of second endotracheal suction and 3) at 120 seconds from the end of second suction.

The number of infants with hypoxemia (defined as Tc PO2 <40 mmHg) and mean Tc PO2 were measured at these points. Other outcomes include Tc PO2 recovery (time taken for Tc PO2 to return to baseline values) and the numbers of infants experiencing bradycardia (heart rate < 100 beats per minute during the suctioning procedure).

For further details see characteristics of included studies.

Two studies were excluded (Cabal 1979 and Graff 1987) as the methods of allocation was unknown and no correspondence has been received from the authors to requests for this information in 2000. One trial (Gonzalez 2005) is awaiting classification pending further information from the authors on the method of allocation to treatment and additional data.