Introduction

In addition to underlying disease, critical illness per se seems to be detrimental to musculoskeletal health in various ways [1,2,3]: immobilization, inflammation, multiple endocrine alterations, hypercatabolism including muscle wasting, malnutrition, and certain drugs all have the potential to disturb the delicate balance between bone formation and bone resorption. Immobilization rapidly induces high bone turnover and bone loss in healthy adults and in patients following stroke with consecutive palsy [4, 5]. Inflammation-induced cytokines like TNF-α and IL-6 are elevated in critical illness [6], may reach extreme levels during sepsis, and induce bone resorption through osteoclastogenesis. Furthermore, hormonal and metabolic changes with adverse effects on bone that frequently affect critically ill patients include hypercortisolism, hypogonadism, hyposomatotropism, and secondary hypothyroidism besides alterations in calcium-phosphate metabolism and parathyroid hormone secretion [7, 8]. Finally, glucocorticoids, proton pump inhibitors, loop diuretics, and catecholamines may have negative effects on skeletal integrity [9,10,11,12]. Vitamin D deficiency also has negative consequences on musculoskeletal health [3].

Data on skeletal consequences of critical care including bone turnover, bone density, and fracture risk are limited. Under normal conditions, bone formation and bone resorption are tightly linked. In critical illness, however, marked uncoupling between osteoblast and osteoclast functions seems to be present [3]. A similar constellation is found in transplantation bone disease and glucocorticoid-induced osteoporosis, both conditions in which bone may be lost very rapidly accompanied by high fracture rates [12,13,14]. Bone resorption markers are high and increase further during intensive care unit (ICU) stay [3, 6, 15, 16] especially in sepsis [17] and following burn injuries [18, 19].

In 2011, a retrospective case-cohort study in 258 women and 481 men reported an increased fracture risk following critical illness; however, vitamin D status was not available in this study [20]. The same group recently demonstrated in a prospective observational study that bone loss in 66 patients was significantly greater in the year following ICU compared to controls from the Geelong Osteoporosis study [21].

More than a decade ago, Greet Van den Berghe et al. reported an effect of “high-dose” intravenous cholecalciferol supplementation (daily vitamin D supplement of ± 500 IU vitamin D3) on bone turnover markers in prolonged critical illness [6]. Vitamin D slightly increased serum osteocalcin (OC) and diminished carboxy terminal propeptide type-I collagen; other bone turnover markers were not affected, however [6]. In patients with arterial hypertension and vitamin D deficiency no significant effect of vitamin D supplementation on bone turnover markers could be observed [22]. In older community-dwelling women, annual oral administration of high-dose cholecalciferol even resulted in an increased risk of falls and fractures [23]. Following heart transplantation, a higher rate of bone loss is associated with lower vitamin D serum levels [14]. Most importantly, vitamin D deficiency is one of the few modifiable risk factors for compromised skeletal health during and following critical illness. In addition, in their scientific opinion paper published in 2016, the European Food Safety Authority (EFSA) stated that “more research is needed to establish the relationship between responses of bone markers (e.g., OC, bone ALP, and urine N-telopeptide crosslinks) to changes in vitamin D status” [24]. Especially in light of the contradictory evidence published on vitamin D supplementation and BMD and fracture risk, more data on faster changing parameters than BMD and fracture development are of interest to detect more subtle effects of vitamin D on bone [24].

We therefore aimed to evaluate the effect of vitamin D supplementation on bone turnover markers in critical illness survivors as a primary outcome in a post hoc analysis of the VITdAL-ICU study, our recent phase III study evaluating the effect of high-dose vitamin D3 on hospital length of stay [25]. Further, we aimed to compare bone mineral density (BMD) and self-reported falls and fractures 6 months after an ICU stay in patients having received placebo versus patients having received vitamin D.

Material and methods

Study design

The detailed study design and methods as well as the results regarding the primary and secondary outcomes including hospital length of stay, length of ICU stay, the percentage of patients with 25-hydroxyvitamin D levels higher than 30 ng/ml at day 7, hospital mortality, and 6-month mortality have been reported previously [25, 26]. In brief, the VITdAL-ICU study was conducted at the Medical University of Graz, a large tertiary academic center in the southeast of Austria with 1538 beds including 123 ICU beds at five intensive care units (1 medical, 1 neurological, 3 surgical units). Adult patients who were expected to stay in the ICU ≥48 h were screened for vitamin D deficiency by measurement of serum 25(OH)D (measured with the IDS-iSYS, an assay based on chemiluminescence technology (Immunodiagnostic Systems, Boldon, UK)).

Major exclusion criteria were severely impaired gastrointestinal motility, pregnant or lactating women, hypercalcemia (total calcium >2.65 mmol/l or ionized serum calcium >1.35 mmol/l), tuberculosis, sarcoidosis, or nephrolithiasis within the prior year.

In accordance with national and EU requirements and the principles of the Declaration of Helsinki [26], written informed consent was waived, delayed, or obtained from a legal surrogate, depending on the circumstances, and was obtained from each patient who regained mental capacity. The study had previously been registered at clinicaltrials.gov (Identifier: NCT01130181).

For the present analyses, only patients who had a 6-month-follow-up—either a personal or a telephone visit—were included.

Randomization and intervention

Patients were randomly assigned to either a “placebo” or “vitamin D” group in a 1:1 ratio, using the “Randomizer for Clinical Trials” tool developed at the Medical University of Graz (www.randomizer.at), stratified according to ICU type and gender. Patients randomized to the vitamin D group received a loading dose of 540,000 IU of vitamin D3 and starting 1 month after the loading dose, five monthly maintenance doses of 90,000 IU oral vitamin D3 or respective placebo. About 90% of patients in both groups did actually receive the monthly doses. All trial participants, investigators, and assessors were blinded. Patients of both treatment arms were allowed to receive standard vitamin D supplements via enteral and/or parenteral nutrition (approximately 200 IE per day) at the discretion of the treating physician.

Outcomes and follow-up

We followed all patients for 6 months, assessing among other variables bone turnover markers as well as the number of falls and fractures. BMD measurements were performed at the lumbar spine and femoral neck (Lunar iDXA®, GE) in the self-selected subgroup of patients who came for a follow-up visit after 6 months. Patients were free to choose personal (n = 80) or telephone (n = 210) 6-month visits based on their preference.

Laboratory analysis

Blood and urine samples were collected on days 0, 3, 7, 28, and where feasible, month 6. OC and ß-Crosslaps (CTX) levels were measured in all 289 patients at baseline and after 6 months in the 80 patients who agreed to a personal follow-up visit at our clinic. CTX levels were measured by electrochemiluminescence immunoassay (Cobas, Roche, Mannheim, Germany). Total serum OC was also measured by electrochemiluminescence immunoassay (Cobas, Roche) according to the manufacturer’s instructions.

Data management

Investigators who were blinded to study-group assignments collected data. Only potential study drug-related adverse events (hypercalcemia, hypercalciuria, mortality, falls, fractures) were monitored and recorded within the 6-month follow-up.

Statistical analysis

Continuous data following a normal distribution are shown as means with standard deviations; parameters with a skewed distribution as medians with interquartile ranges and categorical data are presented as percentages. To compare between the vitamin D and placebo group at baseline, the unpaired Student t-test, the Mann-Whitney-U-test, or the chi-squared test were used. To evaluate changes in bone turnover markers over time and to compare changes in bone turnover markers between the vitamin D and the placebo group, a linear mixed effects model was calculated.

Where appropriate, skewed variables were ln transformed for parametric analyses. A two-sided p-value of less than 0.05 was considered significant. Analyses were performed with SAS, version 9.2, and SPSS, version 22.

Funding

The study was supported by the European Society for Clinical Nutrition and Metabolism (ESPEN), a research grant including provision of study medication from Fresenius Kabi (Germany) and the Austrian National Bank (ÖNB Jubiläumsfonds, Project Nr. 14,143). The funding sources had no influence on the design of the study, data analysis, or manuscript preparation.

Results

Two hundred eighty-nine of the patients recruited between May 2010 and March 2012 [25] had a follow-up visit after 6 months. Eighty patients presented personally in the outpatient clinic after 6 months, the other 209 survivors were evaluated by telephone. Reasons for not returning for a follow-up in person were of geographical nature and the patients’ clinical conditions. Baseline demographic and clinical characteristics of the 289 patients according to vitamin D or placebo were comparable in the two groups (Table 1).

Table 1 Characteristics of the study population with available follow-up data (n = 289)
Full size table

In linear mixed effects models, 25(OH)D showed a significant change over time (p < 0.001) and between the vitamin D and placebo group (p < 0.001). 1,25(OH)D showed a significant change over time (p < 0.001), but not between groups (p = 0.148), as did total calcium (over time: p < 0.001; between groups: p = 0.365), ionized calcium (over time: p < 0.001; between groups: p = 0.242), urinary calcium/creatinine-ratio (over time: p < 0.001; between groups: p = 0.409), as well as CTX (over time: p < 0.001; between groups: p = 0.688)—which decreased—or OC (over time: p < 0.001; between groups: p = 0.972)— which increased. Despite the significant increase, OC remained within normal levels. Sclerostin also showed a significant change over time (p = 0.016), but not between groups (p = 0.140).

There was, however, a significant difference over time as well as between groups for phosphate (over time: p = 0.011; between groups: p = 0.030) and parathyroid hormone (over time: p < 0.001; between groups: p = 0.038) Table 2.

Table 2 Laboratory parameters related to vitamin D, bone, and mineral metabolism of the study population with follow-up data (n = 289)
Full size table

Two clinical fractures occurred in each group within 6 months (spine X-rays were not performed at follow-up). Rates of self-reported falls at 6 months were numerically higher in the placebo group (33/136, 24.3%, in the placebo group vs. 27/153, 17.7%, in the vitamin D group, p = 0.17), this difference was not statistically significant though.

Six months after ICU admission, BMD at the lumbar spine and femoral neck was not significantly different in the vitamin D group compared to the placebo group (Table 3). In the severe and less severe vitamin D deficiency subgroups, there was also no significant difference in BMD between the vitamin D and the placebo group. After 6 months, 42% (33/79) of patients were classified as osteopenic, and 13% (10/79) as osteoporotic by osteodensitometry.

Table 3 Six-month follow-up
Full size table

Discussion

In this post hoc analysis of a prospective RCT in critically ill patients with 25-hydroxyvitamin D levels ≤20 ng/ml receiving high-dose vitamin D3 or placebo, we found no significant differences in bone turnover markers, BMD, fractures, and falls within 6 months after ICU admission between groups. Furthermore, we observed no significant effect of vitamin D supplementation on CTX and OC. There was, however, a significant increase of OC and a significant decrease of CTX after 6 months compared to baseline suppressed OC and elevated CTX in both study groups.

Besides vitamin D deficiency, other detrimental factors affecting skeletal health in this specific population include immobilization, rapid muscle loss during immobilization, inflammation, endocrine alterations, hypercatabolism, malnutrition, and drugs. All these factors likely contribute to increased bone resorption and accelerated bone loss as indicated by elevated CTX. Especially elderly ICU survivors also seem to carry an increased fracture risk, as suggested by the available limited clinical data [20]. This is highly relevant as fragility fractures, especially hip fractures, are deleterious events for an individual leading to decreased quality of life and often the loss of independence even in less vulnerable populations [27,28,29].

To overcome this situation of increased bone resorption and impaired skeletal health, Hollander and Mechanick suggested the consideration of intravenous bisphosphonates to potently decrease bone resorption [2]. They have been studied in small trials in critically ill adults and children and showed promising results [30,31,32]. Two recent retrospective observational studies suggested that preadmission bisphosphonate use in critically ill patients may even be associated with increased survival [33, 34].

The fact that no clear-cut difference was found in BMD after 6 months between patients receiving vitamin D and those receiving placebo might be attributed to the small sample size available for DXA after 6 months and the slow changes in BMD over time. Due to the nature of the investigated population, DXA scans were not performed at baseline. Furthermore, other factors such as immobilization and muscle loss as well as negative energy balances might have a stronger influence on BMD than vitamin D status alone. In view of the aging ICU population [35], the risk of critical illness-related falls and fractures could be substantial, an aggressive multidisciplinary approach may be indicated to reduce morbidity, mortality, and the high costs that have been documented in other, less sick populations [36,37,38].

Our data support the findings by Orford et al. [21], showing increased bone resorption during acute immobilization and improvement after recovery from critical illness. Similar to Orford, our patients showed a significant increase of OC after critical illness back to normal levels, but no commensurate response as a compensatory increase in bone formation.

In addition to limiting muscle loss by avoiding a negative energy balance, treatment of vitamin D deficiency with the aim to reach levels considered necessary for optimal bone health in other populations (above 20 ng/ml) [39] might be one easily adoptable treatment to improve skeletal consequences of prolonged critical illness. In the present study, however, we could not observe a significant effect of vitamin D supplementation on bone turnover markers. The development of suitable—currently unavailable—parenteral vitamin D monopreparations might contribute to more predictable increases in serum 25(OH)D levels [40], as critical illness-associated impaired gastrointestinal function or renal and drug-related compromises of the hepatic CYP450 system necessary for 25-hydroxylation of vitamin D might impair vitamin D uptake [41, 42].

There are important limitations to our study. These are the single center design, the lack of non-Caucasian or pediatric subjects, possibly limiting the generalizability of our findings and the relatively short follow-up for the outcome fractures, which were—as was the occurrence of falls—self-reported. Further, our trial was certainly underpowered for smaller effects, and spine X-rays detecting clinically silent vertebral fractures were not performed. Also, we did not evaluate how many patients were and were not able to ambulate before study inclusion and how many needed walking aids before and after the study, information that might be important regarding the impact of immobilization on bone turnover markers. Lacking information on nitrogen metabolism and balance and its association with bone turnover is another limitation of our study.

In conclusion, we observed increased bone resorption and decreased bone formation in critically ill patients with 25-hydroxyvitamin D levels ≤20 ng/ml, both conditions resolving over a follow-up period of 6 months. Vitamin D supplementation had no significant effect on CTX and OC. However, further studies are necessary to evaluate the relationship between vitamin D supplementation and BTM, BMD, and fracture risk in critically ill individuals with overt vitamin D deficiency. Perhaps a beneficial effect of vitamin D in critically ill patients may only be present in patients with very low vitamin D status, or a beneficial effect of vitamin D on bone health in critically ill patients might be entirely lacking.