Efficacy of nasal irrigation in the treatment of acute sinusitis in children

Wang 2009 reported no statistical difference in symptom scores between baseline and weeks two and three for four nasal and four non-nasal symptoms, except for daytime rhinorrhoea and nocturnal nasal congestion (P < 0.05).

“Wang 2009 compared the baseline mean score from the first week to the second and third week score for four nasal and four non‐nasal symptoms, and reported no statistical difference in symptoms scores, with the exception of daytime rhinorrhoea and nocturnal nasal congestion (P value < 0.05).”

The 2014 update selected one study as meeting the inclusion criteria after reading the full text and included it in the updated review (Wang 2009).

“We only selected one study as meeting the inclusion criteria after reading the full text and included it in the updated review (Wang 2009).”

Wang 2009 reported a statistically significant mean improvement in nasal peak expiratory flow rate for the normal saline group at an undisclosed medium time point and end of study time point, but no raw data were available to report the size of this difference or comment on clinical significance.

“The authors reported a statistically significant mean improvement in nasal peak expiratory flow rate for the normal saline group at an undisclosed medium time point and end of study time point. However, no raw data were available and therefore we cannot report the size of this difference nor comment on the clinical significance.”

Wang 2009 did not blind participants to the nasal spray; for unreported outcomes of sinus X-ray and nasal smear cytology, both outcome assessors were blinded to treatment allocation; blinding was not described for other outcome measures (high risk of bias).

“Wang 2009 did not blind participants to the nasal spray. For the unreported outcomes of sinus X‐ray and nasal smear cytology, both of the outcome assessors were blinded to treatment allocation. Blinding was not described for other outcome measures (high risk of bias).”

In Wang 2009, there was no loss to follow-up, but there are missing data for two participants for the symptom scores used in this review (low risk of bias).

“In Wang 2009 there was no loss to follow‐up. However, there are missing data for two participants for the symptoms scores used in this review. There is one other participant with missing data for an outcome not included in this review (low risk of bias).”

Bollag 1984 and Wang 2009 relied on interpretation and reporting of infant patients' symptoms by parents, contributing to potential bias and limiting interpretation of results.

“Furthermore, Bollag 1984 and Wang 2009 relied on interpretation and reporting of infant patients' symptoms by parents, contributing to potential bias and this is a limitation in the interpretation of the results of these studies.”

Wang 2009 had methodological weaknesses: baseline score for individual and total symptom score was calculated as a mean of daily scores during the baseline period of seven days rather than on the day of entry; there was a statistically significant difference between treatment groups at baseline as well as at two and three weeks after treatment; and there were incorrect data in the tables presented

“The Wang 2009 paper also had methodological weaknesses. Patients recorded daily symptoms in a symptom diary, but the baseline score for individual and total symptom score (TSS) was calculated as a mean of daily scores during the baseline period of seven days, rather than on the day of entry to the trial. We also noted that there was a statistically significant difference between the two treatment”

The two additional studies included since the original systematic review (King 2012; Wang 2009) have not contributed data of sufficient size or quality to materially change the original findings.

“The two additional studies included since the original systematic review have not contributed data of sufficient size or quality to materially change the original findings (King 2012; Wang 2009).”

Wang 2009 reported statistically significant nasal peak expiratory flow rates but presented no data for this outcome, and the authors were unable to provide this (high risk of bias).

“Wang 2009 reported statistically significant nasal peak expiratory flow rates but presented no data for this outcome. The authors were unable to provide this. Wang 2009 also reported significantly improved quality of life scores in the nasal saline group but again no data were presented or available to the review authors for inclusion in this review (high risk of bias).”

One study (Wang 2009) reported a significant difference in quality of life and peak nasal expiratory flow, but significant methodological and reporting flaws limit the interpretation of these data.

“One study, reported a significant difference in quality of life and peak nasal expiratory flow (Wang 2009). However, there were significant methodological and reporting flaws that limit the interpretation of these data.”

In Wang 2009, there were methodological flaws including averaging symptoms over a week including baseline, and from the data the groups could not be concluded equal at baseline nor improvement due to early treatment effect, judged high risk of other bias.

“In addition to the above, there are some methodological flaws, mainly the averaging of symptoms over a week, especially over the first week that included a baseline measurement. We noted that from the data we cannot conclude that the groups were equal at baseline, nor that the improvement was due to an early treatment effect.”

In Wang 2009, not all pre-specified primary outcomes were reported and some outcomes were reported incompletely, with errors in tabulated data, judged high risk of reporting bias.

“Not all of the study's pre‐specified primary outcomes have been reported; some outcomes of interest in the review are reported incompletely In addition, there are errors in the reported tabulated data.”

In Wang 2009, there was a comparatively small loss of participant data (2 out of 69 not included) but no explanation, judged low risk of attrition bias.

“There is a comparatively small loss of participant data (2 out of 69 not included) but no explanation in the paper”

In Wang 2009, most outcome measures were not blinded to participants or researchers, though some outcomes were objective measures less vulnerable to bias, judged high risk of performance and detection bias.

“Poorly described. Most outcome measures were not blinded to participants or researchers. Some outcomes were objective measures, less vulnerable to bias”

In Wang 2009, correspondence from authors confirmed randomisation but no detail on method was provided, judged unclear risk of selection bias.

“Correspondence from authors confirmed randomisation but no detail on method”

In Wang 2009, allocation concealment was not described and there were significant baseline differences between groups, particularly in rhinorrhoea score, judged unclear risk of selection bias.

“Not described in paper, nor obtained from authors. 30 participants assigned to intervention and 39 to placebo group. Significant differences between groups at baseline, particularly in rhinorrhoea score”