Transferability of metabolic balance studies in short bowel syndrome

Metabolic balance studies (MBS) are the gold standard method to assess the intestinal absorptive function in patients with short bowel syndrome (SBS). During a full MBS, patients are admitted to the hospital to collect duplicates of all ingested foods and drinks, as well as their fecal and urinary output, typically over a 72-h period. These collections are further processed to assess absorption of energy, macronutrients (nitrogen, fat and carbohydrate) and electrolytes (sodium, potassium, calcium and magnesium). Full MBS require dedicated laboratory personnel, equipment, knowledge and experience, which explains why they are currently only performed in one center.

We aimed to explore the transferability of full MBS in patients with SBS from the reference center to a clinical center that was new to and unexperienced in MBS.

A collaboration between the centers was initiated to transfer knowledge on how to perform MBS collections, how to process and how to analyze energy, macronutrient and electrolyte content in the collected samples. At practical level, transferability included successful MBS collecting and processing. At analytical level, transferability included the lyophilization of homogenized samples at both centers. The powder that was created at the reference center was measured at both centers to assess the transferability of the analytical methods; while the site-specific powder was measured at the corresponding center to determine full process transferability. The intraclass correlation coefficient (ICC) was calculated to define the absolute agreement between both centers.

A total of 21 MBS were performed in 7 patients with SBS at the new center. A total of 189 samples were collected, equally distributed among combined meals and drinks (n = 63), fecal output (n = 63) and urinary output (n = 63). Meals and drinks samples and fecal output samples were processed in the new center and lyophilized at both centers, with a perfect reliability for dry matter between the centers (ICC = 1). At the level of method transferability, there was an excellent absolute agreement between centers for energy and nitrogen analyses (ICC>0.9). For fat, the mean coefficients of variation (cv) between centers for fecal output (6.3 %) and meals and drinks (12.0 %) were higher, but ICC showed excellent (0.957) to good (0.787) agreement, respectively. Carbohydrate results differed more between the centers (mean cv of 17.5 % for fecal output and 16.4 % for meals and drinks), translating into a moderate (meals and drinks, ICC = 0.654) to good (fecal output, ICC = 0.812) reliability. Similar results and degrees of agreement as obtained for the method transferability were observed with the site-specific powder.

Our study showed the feasibility of a full MBS transferability to a new unexperienced center with promising agreement of the results obtained between both centers. Our findings aim to inspire other specialized intestinal failure centers to adopt the full MBS methodology for both clinical practice and research purposes.