*2.3. Assessment and Classification of Handgrip Strength*

Grip strength data were collected at the time of admission by trained dieticians with a dynamometer (North Coast Medical Exacta™ Hydraulic Hand Dynamometer, North Coast Medical, Inc., 780 Jarvis Drive, Suite 100, Morgan Hill, CA 95037, USA [44]). The unit of measurement was kg, and measurements were performed in a seated position at the edge of the bed using the dominant hand at a 90◦ angle position without contacting any surface [45]. The patients performed three attempts, interrupted by a one-minute break, and the highest result was collected.

#### *2.4. Outcomes*

For this analysis, the primary endpoint was defined as 180-day all-cause mortality. We prespecified additional short-term and long-term secondary endpoints, including adverse clinical outcomes within 30 days (composite endpoint of the original trial including allcause mortality, admission to intensive care unit (ICU), 30-day readmission rate, functional decline, length of hospital stay, non-elective hospital readmission and major complications (including nosocomial infection, respiratory failure, a major cardiovascular event and acute renal failure or gastrointestinal failure during hospitalization)) and activities of daily living assessed by Barthel Index. Further long-term secondary outcomes included QoL and incidence of falls during the 180-days follow-up period. QoL was assessed using: (a) the EuroQol Group 5-Dimension Self-Report Questionnaire (EQ-5D), which ranges from 0 to 1, with higher scores indicating better life quality, and (b) the EQ-5D visual analogue scale (VAS) from 0 to 100, with higher scores indicating better health status.

#### *2.5. Statistical Analysis*

Categorical variables are expressed as counts and percentages, and continuous variables as means and standard deviations. We performed descriptive statistics by calculating mean HGS according to tumor entities (hematological tumors, lung cancer, gastrointestinal tumors, prostate carcinoma, breast carcinoma and others (gynecological cancers, kidney and urothelial cancers, ear, nose and throat carcinoma, genital cancer, skin cancer, pleural mesothelioma and cancer of unknown primary and similar)) and age (10-year intervals) stratified by sex.

The association between sex-specific HGS and clinical outcome was investigated using logistic regression analyses for categorical variables with reporting of odd ratios (ORs) and linear regression for continuous variables with reporting of coefficients (Coef) and 95% confidence intervals (CI). We adjusted the results for important confounders (sex, age, weight, height, NRS 2002 score, center), several main diagnoses (cardiovascular, infectious, renal, frailty), various comorbidities (hypertension, chronic kidney disease, chronic heart failure, diabetes mellitus) and for randomization group.

All statistical analyses were performed with STATA 15.1 (Stata Corp, College Station, TX, USA). A *p* value < 0.05 (for a 2-sided test) was considered to indicate statistical significance.

#### **3. Results**

#### *3.1. Patient Cohort*

Of the initial population of 2028 patients included in the original trial recruiting patients from April 2014 to February 2018, we had complete data for 628 (368 male and 260 female) cancer patients from eight hospitals in Switzerland. The baseline characteristics for all patients included in this analysis, stratified according to sex, are shown in Table 1. Patients had a mean age of 72 years (±12.5), and 41.4% were females. The mean (SD) BMI was 24.6 (±4.8) with a similar distribution in both sexes, and the most common admission diagnosis was cancer (50.8%), followed by infection (21.0%). Patients had a high burden of comorbidities, including hypertension (49.8%), chronic renal disease (30.3%), coronary heart disease (24.8%), diabetes mellitus (19.9%) and chronic heart failure (11.6%). The most frequent types of cancer were hematological tumors (19.7%), lung cancer (16.4%), and gastrointestinal tumors (12.4%).

**Table 1.** Baseline characteristics of malnourished cancer patients.




\* Gynecological cancers, kidney and urothelial cancers, ear, nose and throat carcinoma, genital cancer, skin cancer, pleural mesothelioma and cancer of unknown primary.

#### *3.2. Handgrip Measurement in the Study Population*

The overall mean (SD) HGS was 23.6 (±10.7 kg) with lower values in females (17.3 ± 6.3) compared to males (28.0 ± 10.8). Age, tumor entity and sex-specific HGS data are presented in Table 2. With higher age, the mean (SD) HGS decreased. In younger male cancer patients (<50 years), the mean (SD) HGS was 45.1 kg (±12.7 kg), while in patients ≥ 90 years, there was a mean (SD) HGS of 19.5 kg (±8.0 kg). For female cancer patients, the mean (SD) HGS values were lower, ranging from 23.1 kg (±8.9 kg) in young patients to 8.8 kg (±4.8 kg) in the oldest age group (≥90 years). Stratified by tumor entity, lung cancer patients had the highest mean HGS with 27.4 kg (±10 kg), which was consistent in both sexes (male: mean HGS of 30.8 kg (±9.6 kg), female: mean HGS of 18.9 kg (±4.5 kg)). In the female population, the lowest mean HGS was found in gastrointestinal tumor patients: 16.4 kg (±6.0 kg), whereas male patients had the lowest HGS with prostate carcinoma: 23.6 (±7.4 kg).

**Table 2.** Handgrip strength according to tumor entity and age.


Abbreviations: HGS, handgrip strength; SD, standard deviation. \* Gynecological cancers, kidney and urothelial cancers, ear, nose and throat carcinoma, genital cancer, skin cancer, pleural mesothelioma and cancer of unknown primary.
