Weight Loss in Individuals with Obesity and Asthma.

To the Editor:

We commend Okoniewski and colleagues for their recent systematic review published in AnnalsATS (1) on weight loss interventions in obese subjects with asthma, which we read with great interest. Although not cited by the authors, we note a comparable 2012 Cochrane review by Adeniyi and Young (2), which similarly assessed randomized controlled trials of weight loss interventions in asthma.

We draw your attention to the omission in the present review of at least one study eligible for inclusion (3) according to the a priori inclusion criteria, which did not restrict studies according to language of publication providing sufficient information in English was available for accurate data collection. This study had been published in Spanish but had been included in the aforementioned Cochrane review, for which the data were extracted and published in English. It is notable that the current review, as published, thus departs from its own protocol, registered in the International Prospective Register of Systematic Reviews (PROSPERO CRD42018085045).

We note also the inclusion of a study (4) of a population that combined subjects deemed to be “at high risk of asthma” with subjects with a diagnosis of asthma, despite no disaggregated data being published for these two groups.

Of 10 studies included in the review by Okoniewski and colleagues (1), only that by Stenius-Aarniala and colleagues (5) was published in full at the time of the Cochrane review, and it is striking that both reviews diverge in their assessment of the risk of bias in that study. The study by Stenius-Aarniala and colleagues is assessed in the current review as having low risk of bias in random sequence allocation but was deemed in the Cochrane review to be at unclear risk of bias, as the method used involved shuffling of cards yet produced equal numbers of participants in both groups, despite a small sample size. Conversely, the authors of the present review deemed the risk of attrition bias and reporting bias in this study to be unclear, whereas the authors of the Cochrane review found the Stenius-Aarniala and colleagues study to be at low risk of bias in these domains.

The current review also overlooks biased estimates in its qualitative analysis when referring to one study (6), which reported “a significant improvement in FVC in liters (P = 0.006 between groups), but not FVC % predicted.” It must be noted that the pertinent study used a per-protocol analysis, which did not include those patients randomized to intervention who achieved less than 10% weight loss. The published intent-to-treat analysis included only forced vital capacity (FVC) % predicted, and in this there was no significant between-group difference.

Finally, we must question the decision, in the protocol for the present review, to prioritize “relevant biomarkers (such as leptin, adiponectin, or IL-6)” as a primary outcome, with patient-important outcomes, such as asthma exacerbations, relegated to among the secondary outcomes or not listed at all (e.g., quality of life). These biomarkers were not included as either core or supplemental outcomes by an expert subcommittee convened by the National Institutes of Health and the Agency for Healthcare Research and Quality to propose which biomarkers should be assessed as standardized asthma outcomes in future clinical research studies—neither were they considered to be “emerging” (i.e., requiring validation and standardization) (7).

Notwithstanding these considerations, the review succeeds in highlighting the need for further well-designed randomized controlled trials investigating the effect of weight loss in obese asthma. As more recently approved obesity drugs, including liraglutide and lorcaserin, remain unstudied in this context to date, the time is ripe for such trials.