Morphometric parameters of muscle and bone in critically ill patients

Sarcopenia, characterized as reduction in skeletal muscle mass and strength, can compromise physical function and quality of life. In the literature there are heterogenous definitions of sarcopenia: however, a worldwide widely used definition was proposed in 2010 by the European Working Group on Sarcopenia in Older People (EWGSOP), which takes into account low muscle strength as a key characteristic [1]. Criteria used in other proposed definitions were low skeletal muscle mass either alone or with low muscular function assessed by grip strength and/or slow gait speed [2]. Depending on the criteria used for the definition, the reported prevalence of sarcopenia varies significantly. Based on dual-energy X‑ray absorptiometry (DXA), derived measurements adjusted to body height, the prevalence of sarcopenia was reported to be 53% in men and 43% in women at the age of >80 years [3]. Assessing sarcopenia with standard computed tomography (CT) scans is easy in theory. Millions of CT scans are performed yearly carrying potentially useful information about sarcopenia. Retrieval of data available on body CT examinations performed for any indication requires no additional cost, patient time, equipment or radiation exposure and can also be acquired retrospectively. In the literature, there is heterogeneity about the impact of sarcopenia on mortality. Some published data suggested an independent association of sarcopenia with increased mortality critical illness [4, 5]. Besides total psoas area (TPA), we evaluated qualitative parameters including psoas muscle density (PMD) and bone mineral density (BMD) as well as the skeletal muscle index (SMI).

The objective of the present analysis was to evaluate the ability of these morphometric parameters (TPA, PMD, BMD and SMI) in predicting 6‑month mortality and hospital length of stay (HLOS) in a cohort of mixed vitamin D deficient patients.

This evaluation is a retrospective analysis based on the high dose vitamin D3 vs. placebo in adult vitamin D deficient patients (VITdAL-ICU) clinical trial, which originally included 480 critically ill patients with vitamin D deficiency (defined as a 25-hydroxyvitamin D < 20 ng/ml) [6]. A routine abdominal CT scan for any indication was available in 37 patients between 6 months before and 3 days after intensive care unit admission. A total of 11 CT examinations could not be included in some analyses due to missing scan levels of L3.

The scans were exported from the hospital radiology system as DICOM files. The OsiriX 8.0 software (Pixmeo SARL, Geneva, Switzerland) was used to analyze the scans including measuring and quantifying the morphometric parameters. Clinicopathologic data including age, sex, weight, height, body mass index (BMI), admission diagnosis, survival time and HLOS were available from the original data set. The body surface area was computed using the formula by Mosteller \(\{\text{body surface} [\text{m}^2]= \text{height} [\text{cm}] \times (\text{weight} [\text{kg}]/3600)\}\) . In order to divide the patient cohort into a sarcopenia and non-sarcopenia group (similarly for BMD: likely and unlikely osteoporosis group) we used predefined cut-off values for morphometric parameters based on previous published data [7]. Furthermore, the study group was also separated into survivors (surviving longer than 6 months after ICU admission) and non-survivors (death within 6 months after ICU admission).

The mean age was 58.8 ± 17.3 years, 18 (49%) were women and 19 (51%) were men. The mean BMI value was 26 kg/m² and higher in women than in men. The analyzed population presented a broad variety of admission diagnoses, including neurologic (27%), gastrointestinal (18.9%) and others (Table 1), 57% of the patients were given artificial respiration, 40% were receiving vasopressor agents.

After adjusting to body surface, the mean TPA was 624 mm²/m². When stratified by gender, women showed significantly lower TPA (478 vs. 749 mm²/m², p = 0.002) as well as significantly lower SMI (62.4 vs. 76.8 cm²/m², p = 0.01) than men. The analysis of the other morphometric parameters showed numerically similar results but without statistical significance and are outlined in Table 1. The prevalence of sarcopenia significantly varied depending on the criteria used for definition. Based on TPA, sarcopenia was present in 69% of patients. Defined by PMD and SMI, the percentage of patients presenting with sarcopenia was lower (35% and 4%, respectively). The differences between women and men are summarized in Table 1 and Fig. 5. Using the predefined cut-off values for SMI, there is only one single patient classified as sarcopenic. This should be considered when interpreting the results relating to SMI.

As expected, all parameters showed a strong negative association with age (TPA: r = −0.67, p < 0.001; PMD: r = −0.46, p = 0.01; SMI: r = −0.53, p < 0.003; BMD: r = −0.62, p < 0.001) (Fig. 6).

The mean length of hospital stay was 13 days (interquartile range 7.5–24 days) and did not differ significantly among patients with or without sarcopenia defined by TPA (15.0 vs. 18.1 days; p = 0.49), PMD (13.4 vs. 17.3 days; p = 0.38) and BMD (14.0 vs. 18.6 days; p = 0.34) (Table 2). The 6‑month mortality was similar in patients with and without sarcopenia, irrespective of the morphometric parameter used (Table 3). As mentioned above, the results of SMI in sarcopenia and non-sarcopenia group (hospital length of stay: 31.0 vs. 15.36 days, p = 0.14; mortality rate: 100% vs. 20%, p = 0.52) were not meaningful, as there was just one patient classified as sarcopenic. Assessing TPA, PMD and SMI 11 patients were excluded due to missing scan levels at L3. Assessing BMD one patient was excluded due to vertebral fractures.

Morphometric parameters were not significantly better in survivors (n = 26) compared to non-survivors (n = 11) (TPA: 652 vs. 530 mm²/m², p = 0.27; PMD: 38.4 vs. 37.4 HU, p = 0.85; SMI: 70.3 vs. 69.5 cm²/m², p = 0.91; BMD 151 vs. 146 HU, p = 0.81) (Table 4).

Frailty can be assessed in a number of ways but it appears attractive to use novel but simple analytical methods to assess sarcopenia and bone mineral density with routine CT images. While it was traditionally assessed by only one single cross-sectional measurement of the psoas muscle (TPA) [13], recently several other morphometric parameters (PMD, SMI, BMD) have increasingly been used. Obtaining a predictive marker for prognosis of critically ill patients is of interest as it may improve the risk stratification of critically ill patients.

In our small, retrospective analysis of the VITDAL-ICU trial, we aimed to assess morphometric parameters in all available CT scans and test if there was a strong predictive value of any of these on important clinical outcomes.

All morphometric parameters diminished with increasing age, with the strongest effect noted for TPA. As expected, men showed significantly higher values than women for TPA, PMD, SMI and BMD. These results confirmed that age and sex should be taken into account in the definition of sarcopenia. The morphometric parameters TPA, PMD, SMI and BMD were not significantly associated with both hospital length of stay and 6‑month mortality rate; however, all results were numerically better in survivors. Given that our sample size was very small, we cannot exclude a type 2 error in statistical analysis and that results of a bigger cohort may become significant. Currently, the results in the literature are conflicting [4, 14‐16]. Notably, we did not evaluate the influence of volumetric parameters (i.e., total psoas volume), that have been shown to be associated with patient mortality and hospital length of stay [17]. Several limitations of the present analysis should be considered. Due to the low number of patients, results should be interpreted with caution. Also, CT examinations were not performed at a specific time point due to the retrospective nature of this study. For the same reason, it was not possible to measure other frailty parameters, for example grip strength. Because of the use of routine CT examinations, unenhanced and contrast-enhanced images were both analyzed. Thus, errors in measuring attenuation values could not be excluded. Furthermore, comorbidities such a malignant diseases or medication that may affect bone density were not assessed and may therefore distort results.

Defining sarcopenia obtained by morphometric parameters in CT scans is also unable to include muscle function. As there is no linear connection between muscle mass and muscle function [18], both should be included for the diagnosis of sarcopenia. As this is, however, impossible in ventilated or severely ill patients, the identification of a morphologic surrogate parameter could be of interest.

In conclusion, sarcopenia assessed with morphometric parameters obtained from CT scans was not an independent predictive factor of hospital length of stay and 6‑month mortality in critically ill, vitamin D deficient patients. The validity of these results is limited due to the relatively small patient cohort of 37.