Increased fat accumulation may be associated with severe muscle wasting in critically ill patients: a prospective observational study

Abstract

Obesity has been hypothesized to attenuate muscle wasting in critically ill patients due to increased ketogenesis. This study examined the associations between fat mass volume, ketone bodies, and muscle wasting in critically ill patients. We conducted a prospective study in an emergency intensive care unit (ICU) from November 2021 to October 2023, enrolling adult patients with an expected ICU stay of ≥ 7 days and abdominal computed tomography (CT) performed within 24 h of admission. Patients were classified as adipose or lean based on fat area measured via CT. The primary outcome was severe muscle wasting, defined as a > 10% decrease in the rectus femoris cross-sectional area measured by ultrasonography from day 1 to day 7. Among 134 enrolled patients, 108 were evaluable (57% male, median age 73 years). Severe muscle wasting was more frequent in the adipose group (48%) than in the lean group (27%, p = 0.023). Multivariate analysis confirmed a higher risk of muscle wasting in the adipose group (adjusted OR 2.52, p = 0.034). BHB levels were inversely correlated with fat area and associated with a reduced risk of muscle wasting. Contrary to our hypothesis, obesity increased the risk of muscle wasting, potentially due to decreased ketogenesis.

200/200 words

Introduction

Persistent dysfunctions after discharge from the Intensive Care Unit (ICU) are a significant issue for ICU survivors, encompassing cognitive, psychological, and physical impairments. Physical impairment most commonly develops as ICU-acquired weakness (ICU-AW), characterized by generalized and symmetrical muscle weakness secondary to critical illness. Muscle wasting is a typical feature of ICU-AW, with skeletal muscle mass decreasing by nearly 2% per day during the first 7 days of the ICU stay¹.

Obesity is characterized by excessive fat accumulation and is generally associated with increased risk of several diseases and mortality². Conversely, protective effects of obesity have been reported in several conditions, including critical illness³, and this phenomenon is called the “obesity paradox”. The protective effect of obesity was also suggested against muscle wasting and ICU-AW, and increased lipolysis and ketogenesis in obese individuals may mediate this protective effect^4,5. However, in most previous clinical studies evaluating the obesity paradox, obesity was defined solely based on excessive body mass index (BMI)⁶. Because BMI reflects not only fat tissue but also other lean tissues, such as skeletal muscles, it remains unclear whether these protective effects originate from increased fat accumulation. Therefore, evaluating the direct relationship between muscle wasting and fat accumulation is crucial.

We hypothesized that increased fat accumulation attenuates muscle wasting during ICU stay due to increased lipolysis and ketogenesis. Because several functions of adipose tissue differ according to its distribution (i.e., subcutaneous and visceral fats), we also hypothesized that the association between fat accumulation and muscle wasting is affected by fat distribution. To evaluate this hypothesis, we conducted a single-centre prospective cohort study.

Materials and methods

Study setting and population

We conducted a single-centre prospective cohort study in a 20-bed emergency ICU at Yamaguchi University Hospital. The study was approved by the Institutional Review Board of Yamaguchi University Hospital and was conducted in accordance with the Declaration of Helsinki and the Ethical Guidelines for Medical and Health Research Involving Human Subjects in Japan. Adult patients admitted to the ICU between November 2021 and October 2023 with an expected ICU stay of 7 days or longer and abdominal computed tomography performed within 24 h of ICU admission were screened for eligibility. Patients aged < 18 years, patients whose abdominal CT was inadequate for measurement of muscle and fat, patients with premorbid severe frailty (i.e., clinical frailty scale score ≥ 7), patients who suffered from neuromuscular diseases or stroke, patients with bilateral femoral injuries that prevented muscle ultrasonography, and other patients who were considered inadequate for the study by their attending physicians were excluded. There was no specified upper age limit. Patients with COVID-19 were also excluded because of the restraint for some study measurements. Eligible patients were included in the present study if written consent could be obtained from patients or their proxies.

In our institution, because there is no documented protocol for nutritional therapy or rehabilitation, they are individually prescribed by their attending physicians. In general, tube feeding via a gastric tube or oral feeding was initiated if patients were hemodynamically stable with or without vasopressor use and had no contraindication for enteral feeding. During the first week, underfeeding was permitted, and parenteral nutrition or liquid nutrients were not routinely prescribed. Supplemental protein, in addition to standard enteral and parenteral formulas, was not applied. The rehabilitation therapies provided by physiotherapists were provided at the appropriate intensity to each patient once or twice a day on weekdays. ICU nurses also provided some rehabilitation therapies.

Data collection

We collected the following baseline data: age, sex, height, body weight, BMI, clinical frailty scale score, premorbid activity of daily living (ADL) score which was evaluated by the Barthel index and functional independence measure (FIM), cause of ICU admission, Acute Physiology and Chronic Health Evaluation (APACHE) II score, and Sequential Organ Failure Assessment (SOFA) score on admission. The Barthel index and FIM were assessed via interviews with patients, family, or their caregivers. The SOFA score was also evaluated on days 3 and 7. On days 1, 3 and 7, blood samples were taken for measurements of ketone bodies (i.e., beta hydroxybutyrate [BHB] and acetoacetate [AA]), lipid profiles (i.e., free fatty acids [FFAs], triglycerides, low/high-density lipoprotein cholesterol [LDL/HDL-C]), and adipokines (i.e., adiponectin and leptin). Adiponectin and leptin were analysed via commercially available enzyme-linked immunosorbent assay kits (R&D Systems; MN, USA). The mean number of calories and proteins applied, and the amounts of some drugs (insulin, catecholamines, neuromuscular blockers, and systemic steroids) were calculated up to day 7. During the ICU stay, the use of mechanical ventilation and renal replacement therapy were recorded.

In this study, we used abdominal CT to estimate the amount of fat tissue or skeletal muscle. Using abdominal CT performed within 24 h of ICU admission, we calculated the fat area and abdominal muscle area (AMA) according to previous studies⁷. In detail, the fat area and AMA at the level of the third lumbar spine were measured on an image-processing workstation (AZE VirtualPlace Raijin; AZE Ltd, Tokyo, Japan). The fat area was calculated as the area with a Hounsfield unit (HU) range of − 190–30 and divided into the visceral fat area (VFA) and the subcutaneous fat area (SFA) semiautomatically by the preinstalled program of the workstation. The total fat area (TFA) was defined as the sum of VFAs and SFAs. The AMA was calculated as the area with − 29 to + 150 HU within the manually outlined skeletal muscles, including the psoas, erector spinae, quadratus lumborum, transversus abdominis, external and internal oblique abdominal muscles, and rectus abdominis muscle. In addition, the quadriceps femoris muscle area (QMA) was measured in the same way just below the lesser trochanter of the femur on both sides, and the mean value was used for analyses. These body composition analyses were also performed on subsequent abdominal CT scans performed between days 4 and 10, if available.

Outcomes

To evaluate muscle wasting during the ICU stay, we measured the cross-sectional area of the rectus femoris muscle (CSA_RF) via ultrasonography on days 1, 3, and 7, according to previous studies⁸. Muscle ultrasonography was performed by experienced ICU physicians blinded to patient grouping. At the time of CSA_RF measurement, patients were placed in the supine position with their limb extended and relaxed. The CSA_RF was measured at the midpoint between the anterior superior iliac spine and the proximal border of the patella, and the markings were provided at the initial measurements for subsequent measurements. For the measurement, a linear echo probe was placed perpendicularly to the rectus femoris muscle with minimal compression pressure and excessive echo gel to avoid tissue distortion. Each measurement was performed 3 times on each side, except for patients with unilateral limb injury preventing muscle ultrasonography, and the CSA_RF was calculated as the mean value. We determined the primary outcome in this study as severe muscle wasting, which is defined as a decreased CSA_RF greater than 10% between days 1 and 7⁸ because other parameters that directly reflect muscle weakness (e.g., handgrip strength and the Medical Research Council [MRC] sum score) could rarely be compared between pre- and postmorbid states in cases of unplanned admission. We also evaluated the following secondary outcomes: changes in CSA_RF from day 1 to days 3 and 7; changes in body composition (i.e., FA, AMA, and QMA) between baseline and follow-up CTs; handgrip strength and MRC sum score at ICU discharge; survival at hospital discharge; and changes in the Barthel index and FIM between premorbid and hospital discharge. Changes in body composition between two CTs were evaluated as percentages of changes per day, which were calculated as differences from CT on admission to follow-up CT divided by the baseline value and the number of days between two CTs. Survival, the Barthel index, and FIM were also evaluated by telephone interviews to patients, families, or their caregivers at 6 and 12 months. Patient-reported quality of life (QOL) was assessed via mailed 36-item Short Form health survey questionnaire (SF-36) version 2 at 6 and 12 months.

Statistical analysis

The variables are shown as medians (interquartile ranges) or numbers (percentages) of subjects. We divided patients into the adipose and the lean groups based on TFA. Since no specific cutoff values for TFA have been established, we used the median value of TFA for group division. The background characteristics, laboratory data, and outcomes were compared between the groups. Univariate analyses were performed with the χ² test for categorical variables and the Mann–Whitney U test for continuous variables. Correlations between continuous variables were assessed using the Spearman correlation coefficient. Additionally, the adjusted odds ratio (OR) for severe muscle wasting in the adipose group was calculated via multivariable logistic regression analysis. To determine the association between fat distribution and muscle wasting, we also calculated the adjusted ORs of the subcutaneous and visceral adipose groups, which were determined by the median values of SFA and VFA, respectively. The variables used for adjustment included age, sex, and APACHE II score. Sensitivity analyses were performed by adding some additional variables (i.e., admission diagnosis, mechanical ventilation, CSA_RF on day 1, mean daily applied calories or proteins during the first week, and total fluid balance during the first week) to the ordinary models. We also calculated the adjusted ORs for BHB, leptin, and adiponectin on day 1, and CRP on day 3, in the same way. For all tests, two-tailed p values of < 0.05 were considered statistically significant. We used IBM SPSS version 19 (IBM SPSS Inc., Chicago, IL) for all the statistical analyses.

Although clinical incidences of severe muscle wasting in the obese and lean patients were unknown, one retrospective cohort study reported that the incidence of ICU-AW, which was defined as an MRC sum score < 48, was 12% in the obese group and 24% in the lean group⁴. Additionally, the incidence of severe muscle wasting during the first 7–10 days was previously reported to be 50–66%^9–13. Therefore, we supposed that the overall incidence of severe muscle wasting was 50% and the incidence in the obese group was approximately half of that in the lean group (i.e., 35% in the obese group and 65% in the lean group, respectively). Based on these predicted incidences, we performed a power analysis with a statistical power of 80% at the 5% significance level and estimated a minimal sample size of 42 patients in each group. Considering the expected unevaluable primary outcome due to early ICU discharge or death, we planned a total of 100 inclusions in the present study.

Results

During the study period, 395 adult patients with abdominal CT data were screened (Fig. 1), and 134 patients (33.9%) were included. Among these patients, the primary outcome was not evaluable in 26 patients (19.4%), mainly because of an ICU stay shorter than 7 days, including 3 patients who died within 6 days; these patients were excluded from the analyses. The background characteristics and treatments of the analysed patients are shown in Table 1. Among the analysed patients, 62 patients (57%) were male, and the age and APACHE II score were 73 (59–80) years and 19 (13–28), respectively. The BMI of the analysed patients was 22.9 (19.6–25.8) kg/m², and 15 (14%), 20 (19%), and 34 (31%) patients were classified as underweight (< 18.5 kg/m²), overweight (23–24.9 kg/m²), or obese (> 25 kg/m²), respectively, according to the Asia–Pacific classification¹⁴. Although frailty or premorbid ADLs did not differ significantly between the two groups, the adipose group had significantly larger muscle areas, BMIs and fat areas than did the lean group. Among the analysed patients, 39 patients (36%) and 12 patients (11%) were treated with mechanical ventilation and renal replacement therapy, respectively. Enteral nutrition was initiated within 48 h after admission in 64 patients (59%), and the mean daily intake of calories and proteins during the first week was 650 (438–827) kcal/day and 24.5 (16.1–33.3) g/day, respectively. When nutritional doses were calculated on the basis of actual body weight for patients with a BMI ≤ 25 and ideal body weight for those with a BMI > 25, as recommended in the current guidelines¹⁵, the adipose group tended to receive lower caloric (10.2 [8.3–14.6] kcal/kg/day vs 13.0 [9.2–17.1] kcal/kg/day, p = 0.062) and protein (0.40 [0.27–0.61] g/kg/day vs 0.52 [0.32–0.68] g/kg/day, p = 0.093) doses, although these differences were not statistically significant (Table 1 and Supplementary Fig. 1). There were no significant differences in the other treatments between the two groups, except for the higher insulin dose in the adipose group.

Fig. 1.

Patient flowchart.

Table 1.

Background characteristics and treatment.

	Adipose group	Lean group	p value
	(n = 56)	(n = 52)
Age (years)	70 (59–77)	76 (59–84)	0.074
Sex (male)	32 (57)	30 (58)	0.954
BMI (kg/m2)	25.3 (23.6–28.3)	19.7 (18.3–22.2)	< 0.001
Charlson comorbidity index	1 (0–2)	1 (0–2)	0.607
Clinical frailty scale	3 (2–3)	3 (2–4)	0.275
Pre-morbid ADL
Barthel index	100 (100–100)	100 (100–100)	0.099
FIM	126 (124–126)	126 (118–126)	0.708
Admission diagnosis			0.827
Sepsis	9 (16)	7 (13)
Respiratory failure	5 (9)	4 (8)
Other medical conditions	20 (36)	19 (37)
Trauma	19 (34)	16 (31)
Other non-medical conditions	3 (5)	6 (12)
APACHE II score	21 (13–28)	18 (13–26)	0.620
SOFA score	5 (2–9)	4 (2–7)	0.553
Body compositions measured by abdominal CT
Total fat area (cm2)	357.8 (298.8–433.6)	122.0 (89.1–180.5)	< 0.001
Subcutaneous fat area (cm2)	184.3 (142.9–198.6)	54.4 (41.1–79.2)	< 0.001
Visceral fat area (cm2)	180.0 (146.5–240.6)	63.0 (46.2–97.2)	< 0.001
Abdominal muscle area (cm2)	120.5 (96.4–143.4)	100.1 (90–119.9)	0.002
Quadriceps femoris muscle area (cm2)	38.3 (29.5–44.1)	29.4 (23.8–35.9)	0.001
Treatments
Drugs used during first week
Insulin	24 (43)	10 (19)	0.008
Cumulative dose during first week (unit) *	29 (5–100)	19 (12–27)	0.564
Catecholamine	16 (29)	14 (27)	0.848
Cumulative dose during first week (μg/kg) *†	149.9 (66.6–483.3)	164.5 (47.8–400.6)	0.918
Rocuronium	3 (5)	3 (6)	0.926
Systemic corticosteroid	11 (20)	13 (25)	0.503
Mean dose during first week (mg/day) *‡	36.4 (26.6–47.8)	26.1 (14.3–92.9)	0.608
Total fluid balance during first week (ml)	1509 (-1439–4566)	1412 (-1095–4169)	0.917
Mechanical ventilation	24 (43)	15 (29)	0.130
Renal replacement therapy	7 (13)	5 (10)	0.634
Nutrition during first week
Early enteral nutrition	32 (57)	32 (62)	0.642
Mean daily applied calories (kcal/kg/day) §	10.2 (8.3–14.6)	13.0 (9.2–17.1)	0.062
Mean daily applied proteins (g/kg/day) §	0.40 (0.27–0.61)	0.52 (0.32–0.68)	0.093

* Only patients to whom the drug was applied were analyzed.

^† Shown as noradrenaline dosages which calculated using dosing equivalencies.

^‡ Shown as prednisolone dosages which calculated using dosing equivalencies.

^§ Nutritional doses were calculated based on actual body weight for patients with BMI ≤ 25 and ideal body weight for those with BMI > 25, as recommended in the current guidelines¹⁵.

ADL, activities of daily living; APACHE, Acute Physiology and Chronic Health Evaluation; BMI, body mass index; FIM, functional independence measure; SOFA, Sequential Organ Failure Assessment.

The primary and secondary outcomes are shown in Table 2 and Fig. 2. As shown in Fig. 2A, although CSA_RF was significantly greater in the adipose group throughout the first week, the loss of CSA_RF by day 7 was significantly greater in the adipose group than in the lean group (-9.3 [-14.8–-4.5] % vs -4.5 [-10.4–1.3] %, p = 0.014; Fig. 2B). The change in CSA_RF by day 7 had weak but statistically significant negative correlations with TFA, SFA, and VFA (Supplementary Fig. 2). The prevalence of severe muscle wasting was significantly greater in the adipose group (27 [48%]) than in the lean group (14 [27%], p = 0.023). Among the 63 patients (58%) whose follow-up CT data were available for measurements, the loss of QMA was also significantly greater in the adipose group than in the lean group (-1.2 [-1.9–-0.8] %/day vs -0.8 [-1.3–0.6] %/day, p = 0.005), and the loss of AMA tended to be greater (-0.44 [-0.89–0.20] %/day vs -0.17 [-0.73–1.18] %/day, p = 0.063). Changes of neither SFAs nor VFAs significantly differed between the two groups. The length of ICU stay was significantly longer in the adipose group than in the lean group. Handgrip strength and MRC sum score at ICU discharge were evaluable in 97 patients and 100 patients, respectively, and were significantly greater in the adipose group. The proportion of patients whose MRC sum score was less than 48 at ICU discharge tended to be lower in the adipose group (13%) than in the lean group (25%), but the difference was not statistically significant. CSA_RF, especially which measured on day 7, was significantly correlated with the handgrip strength and MRC sum score at ICU discharge (Supplementary Fig. 3). Survival rate, length of hospital stay and ADLs (i.e., the Barthel index and FIM score and their changes from the premorbid state) at hospital discharge did not significantly differ between the two groups. Moreover, long-term survival, QOL and ADL were not significantly different between the two groups (Supplementary Table 1).

Table 2.

Primary and secondary outcomes.

	Adipose group	Lean group	p value
	(n = 56)	(n = 52)
Severe muscle wasting *	27 (48)	14 (27)	0.023
Change of CSARF (%)
by day 3	-3.2 (-6.6–0.9)	-0.7 (-8–6.5)	0.124
by day 7	-9.3 (-14.8–-4.5)	-4.5 (-10.4–1.3)	0.014
Change of body composition between CTs (%/day), n = 63
Subcutaneous fat area	-0.27 (-0.81–0.33)	-1.04 (-2.75–1.12)	0.214
Visceral fat area	-0.27 (-0.83–1.54)	0.14 (-1.56–2.48)	0.717
Abdominal muscle area	-0.44 (-0.89–0.2)	-0.17 (-0.73–1.18)	0.063
Quadriceps femoris muscle area	-1.2 (-1.9–-0.8)	0.8 (-1.3–0.6)	0.005
Muscle strength at ICU discharge
Handgrip strength (kg), n = 97	22.5 (14.1–29.6)	15.1 (9.3–26.2)	0.036
MRC sum score, n = 100	59 (53–60)	54 (48–58)	0.002
Less than 48 points	7 (13)	12 (25)	0.142
Length of ICU stay (day)	13 (9–19)	11 (8–15)	0.042
Length of hospital stay (day)	24 (15–36)	27 (14–43)	0.623
In-hospital survival	55 (98)	51 (98)	0.958
Hospital discharge to home	16 (29)	15 (29)	0.975
ADL at hospital discharge †
Barthel index	70 (25–90)	55 (25–100)	0.993
Change from pre-morbidity	-35 (-65–-10)	-25 (-70–0)	0.673
FIM	93 (62–119)	86 (49–123)	0.853
Change from pre-morbidity	-30 (-63–-7)	-27 (-71–0)	0.931

* Defined as loss of CSA_RF by day 7 over 10%.

^†Barthel index and FIM in non-survivors were scored as 0 point.

ADL, activities of daily living; CSA_RF, cross-sectional area of rectus femoris; FIM, functional independence measure; MRC, medical research council.

Fig. 2.

CSA_RF and its change during the first week in the adipose and lean groups, The absolute CSA_RF value (A) and its change from baseline (B) are shown until day 7. The data are presented as median values and interquartile ranges. The black dots and grey triangles indicate the adipose and lean groups, respectively. CSA_RF, cross-sectional area of the rectus femoris, * p < 0.05.

The adjusted ORs of severe muscle wasting for adiposity are shown in Fig. 3. The adipose group was independently associated with an increased risk of severe muscle wasting (adjusted OR 2.52; 95% confidence interval [CI] 1.07–5.90, p = 0.034). When either the subcutaneous or visceral adipose group was included in the model instead of the overall adipose group, the subcutaneous adipose group remained significantly associated with an increased risk of severe muscle wasting (adjusted OR: 3.02; 95% CI: 1.27–7.18; p = 0.013), whereas the visceral adipose group was not (adjusted OR: 2.01; 95% CI: 0.86–4.70; p = 0.106). In the sensitivity analyses, these results were not influenced by adding additional variables (e.g., CSA_RF on day 1, mean daily applied calories or proteins during the first week) to the models, except for mechanical ventilation (Supplementary Table 2). When mechanical ventilation was added, although the association between the total adipose group and severe muscle wasting was not significant, the subcutaneous adipose group remained independently associated with an increased risk of severe muscle wasting.

Fig. 3.

Association of adipose group with the risk of severe muscle wasting, The models were adjusted for age, sex, and Acute Physiology and Chronic Health Evaluation II score at admission.

The results of the biochemical analyses are shown in Supplementary Table 3. BHB on day 1 tended to be lower in the adipose group (156 [76–313] µmol/L vs 252 [109–569] µmol/L, p = 0.081), and BHB on day 1 had a weak but significant negative correlation with TFA (r = -0.221, p = 0.021; Supplementary Fig. 4). Despite the positive correlation between BHB on day 1 and disease severity indicated by APACHE II score, a significant negative correlation was shown between BHB on day 1 and the change in CSA_RF by day 7 (r = -0.262, p = 0.006; Supplementary Fig. 4). In the multivariate logistic regression analysis (Supplementary Table 4), elevated BHB on day 1 was independently associated with a decreased risk of severe muscle wasting (adjusted OR per 100 µmol/L 0.91; 95% CI 0.83–0.99, p = 0.036). BHB on days 3 and 7 and FFAs during the first week did not significantly differ between the two groups. Although significantly higher leptin and lower adiponectin levels were detected in the adipose group throughout the first week, these adipokines were not independently associated with the risk of severe muscle wasting (Supplemental Fig. 5). C-reactive protein (CRP) on day 3 and the leukocyte count on day 7 were significantly greater in the adipose group. CRP on day 3 was also significantly correlated with TFA and the loss of CSA_RF by day 7 (Supplementary Fig. 6) and was associated with the risk of severe muscle wasting (adjusted OR per 1 mg/dL 1.07; 95% CI 1.02–1.13, p = 0.007; Supplementary Table 4).

Discussion

In the present study, we evaluated muscle wasting during the first week of the ICU stay and compared it between adipose and lean patients. In contrast to our hypothesis, severe muscle wasting was significantly more common in adipose patients, particularly those with increased subcutaneous fat accumulation, than in lean patients. To the best of our knowledge, this is the first prospective study to evaluate the association between muscle wasting during ICU stay and the amount of adipose tissue estimated by abdominal CT.

The paradoxical protective effect of obesity, often referred to as the ‘obesity paradox’, has been reported in several contexts, including muscle wasting and ICU-AW^4,5. However, most previous studies that evaluated the “obesity paradox” defined obesity based on BMI. Although BMI is a standard indicator for diagnosing obesity, it may not always accurately reflect the extent of fat accumulation, as BMI can be influenced by both fat and lean mass volumes, including skeletal muscle. A previous study revealed a low sensitivity (39.6%) of BMI for high fat accumulation measured by bioimpedance spectroscopy, and misclassification as nonobese was more common in patients with a smaller skeletal muscle volume¹⁶. Therefore, obesity should be defined based on the extent of fat accumulation assessed by some methods to determine whether increased fat accumulation itself could impact clinical outcomes.

Previous studies that examined the “obesity paradox” in muscle wasting of critically ill patients reported inconsistent results^4,17,18. Goossens et al. reported that lean critically ill patients had smaller myofibers compared to lean healthy controls, but this was not observed in obese critically ill patients compared to obese healthy controls⁴. However, intraindividual changes in muscle parameters could not be evaluated in this post hoc analysis of their randomized control study. A prospective study by Segaran et al. revealed that muscle depth changes measured by ultrasonography did not significantly differ between lean and overweight/obese patients, but this study included a very small number of patients¹⁷. Additionally, a separate prospective study investigating risk factors for severe muscle wasting found that BMI was not an independent risk factor¹³. However, none of these studies evaluated the relationship between muscle wasting and fat accumulation itself. Only one retrospective study of COVID-19 patients compared obese and lean patients based on the extent of fat accumulation estimated by chest CT¹⁸. In this study, more severely decreased cross-sectional areas of the erector spinae muscles were observed in obese patients than in nonobese patients, which is consistent with our results. However, the timing of CT scanning used for muscle evaluation in this study varied among patients because of its retrospective nature, and this difference may affect the extent of muscle wasting evaluated. In contrast to these previous reports, we defined obesity based on fat volume and prospectively evaluated changes in muscle mass at predefined time points.

At the time of planning the present study, we hypothesized that the risk of severe muscle wasting was decreased in the adipose group because of their higher BHB levels. BHB is reported to have several protective effects (e.g., anti-inflammatory, antioxidative, mitochondria-protective and neuroprotective effects)¹⁹, and it may improve muscle wasting by reducing assumed mechanisms of muscle wasting in critically ill patients with these effects. Indeed, exogenous BHB administration has been suggested to reduce muscle damage in septic mouse models^20,21 and lipopolysaccharide-injected humans²². In the present study, increased BHB on day 1 was also independently associated with a lower risk of severe muscle wasting, as expected. Because endogenous BHB is synthesized mainly from acetyl-CoA supplied from lipolysis, increased BHB in obese patients was expected, as shown in a mouse model⁴. However, BHB during the acute phase tended to be lower in the adipose patients and had a negative correlation with fat volume, in contrast to our hypothesis. This unexpected decrease in BHB in the adipose group during the acute phase may explain their increased risk of severe muscle wasting. Comparable serum FFA levels between the adipose and lean groups suggested that lipolysis occurred equally in both groups, and ketogenesis may be more suppressed in the adipose group for several reasons, such as a higher insulin level.

We hypothesized several other mechanisms for the increased risk of severe muscle wasting in the adipose group in the present study. First, stronger acute inflammation in the adipose group may lead to an increased risk of severe muscle wasting. Although chronic inflammation and its contribution to the onset of metabolic syndrome in obese patients are well documented³, the influence of obesity on the acute inflammatory response in critically ill individuals remains controversial²³. In the present study, a stronger acute inflammatory response in the adipose group was suggested by higher CRP level on day 3. Elevated CRP levels on day 3 were also associated with an increased risk of severe muscle wasting. Because early systemic inflammation is known as one of the major risk factors for muscle wasting¹², this stronger inflammation is considered a possible cause of the increased risk of severe muscle wasting in the adipose group. Second, premorbid muscle wasting may attenuate further muscle wasting during ICU stay and may be considered one of the mechanisms for the lower incidence of severe muscle wasting in the lean group. However, the significant association between adiposity and increased risk of severe muscle wasting remained, even after adjusting for CSA_RF on day 1. Third, differences in the applied doses of calories and proteins between groups might influence muscle wasting. In the present study, because the total amount of applied nutrients did not differ between the groups, the higher actual body weight in the adipose group resulted in lower nutritional doses in this group. These lower nutritional doses might lead to more severe muscle wasting. However, it is still uncertain how many calories or proteins should be applied during the acute phase of critical illness, especially for underweight or obese patients²⁴, and higher nutritional doses did not improve muscle strength or QOL in recent interventional studies^25,26. In the present study, the adjustment for applied calories or proteins did not affect the association between adiposity and the increased risk of severe muscle wasting.

The present study had several limitations. First, this was a single-center cohort study that included only Japanese patients. In addition to institutional bias, such as in nutrition, rehabilitation, and patient characteristics, the racial uniformity of patients could impact the results. Fat accumulation is strongly influenced by races, and Asian populations, including Japanese people, usually have lower body weights and fat volumes than other races do. Additionally, Asian populations are considered to have different associations with the extent of obesity and health risks than other populations are²⁷. Therefore, our results should be externally validated. Second, an adequate diagnostic method of obesity based on the extent of fat accumulation assessed by CT and its thresholds, especially for TFA and SFA, remain uncertain. However, the threshold for VFA to define “visceral obesity”, which indicates excess visceral fat accumulation associated with an increased risk for metabolic syndrome, in Asian populations has been proposed to be approximately 100 cm^228,29. Considering the distributions of VFA in the adipose and lean groups, which bordered around this value, we believe that each group predominantly represented obese and nonobese Japanese patients, respectively. Third, severe muscle wasting did not reflect clinically important outcomes such as muscle strength. In contrast, the prevalence of muscle weakness defined based on the MRC sum score was lower in the adipose group than in the control group, similar to the findings of a previous study⁴. Because the muscle weakness observed in critically ill patients may be affected not only by muscle and other (e.g., polyneuropathy) abnormalities that develop during the ICU stay but also by premorbid states (e.g., sarcopenia), the extent of muscle wasting may not represent this muscle weakness. Indeed, CSA_RF in the adipose group was significantly greater than that in the lean group even on day 7, despite greater loss of CSA_RF, because of their greater muscle volume at baseline. Although changes in muscle strength between the pre- and postmorbid periods should be assessed to evaluate “ICU-acquired” muscle weakness, premorbid muscle strength is difficult to assess in cases of unplanned admission. Additionally, evaluating muscle quality using sonographic measures such as muscle echo intensity and pennation angle, which were not assessed in this study, may provide further insights into predicting these outcomes. Fourth, we evaluated muscle wasting only until day 7 of the ICU stay. Therefore, it is unclear whether increased fat accumulation also affects muscle wasting beyond the early phase of critical illness.

In the present study, adipose patients had a greater risk of severe muscle wasting during the early phase of their ICU stay than lean patients did. Decreased early ketogenesis in adipose patients may mediate this increased risk. These results suggest that further studies are needed to evaluate the effect of ketone supplements on the prevention of muscle wasting and ICU-AW, especially in patients with increased fat accumulation.

Author contributions

Conceptualization: YK, MF. Formal analysis: YK, KK. Investigation: YK, KH, MS, RA, TI. Writing—Original Draft: YK, MF. Writing—Review & Editing: KH, MS, RA, TI, KK, RT.

Funding

This study was supported by a Grant-in-Aid for Scientific Research (21K09076).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

The study was approved by the Institutional Review Board of Yamaguchi University Hospital (approval number 2021–132-2). Written informed consent was obtained from all patients or their proxies.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-96171-8.