Table 2.
Main characteristics of the included studies.
| First Author, Year | Study Details (Study Type, Participants, Years, Country) | Intestinal Failure/pn- Dependence Definitions | Bone Disease Measurement and Definitions | Study Population (Number, Age, Sex, Length of pn) | Control Group or Reference Population | Prevalence of Bone Disease and/or Comparison with Control Group | Vitamin d Status | Follow-Up Results | Factors Associated with Bone Density or Metabolic Bone Disease |
|---|---|---|---|---|---|---|---|---|---|
| Cannon RA, 1980 [16] | Observational, single-center study (OSCS). Children on HTPN at August 1979 with a follow-up of 2 years. USA. |
TPN begun within 2 months of age and still ongoing. | Development of pathologic fractures or loose teeth plus radiologic evidence of rickets were assessed. | 8 children, age 7–24 months, 5 males. Duration of TPN: 7–24 months. |
No | 37.5% had bone disease. | Supplementation with vitamin D resulted in healing of bone lesions in one patient. | Not evaluated | Not reported |
| Larchet M, 1991 [17] | OSCS. Children on cyclic HPN at November 1983, with a follow-up of 4 years. France. |
HPN for SBS. | Osteopenia assessed on wrist and tibia X-ray. Sserum Ca, P, ALP, PTH, and urinary Ca and P were annually measured. Bone biopsy. |
7 children, age 4–14 years, 4 males. Duration of PN at study start: 19–79 months. |
No | 85.7% had signs of moderate osteopenia. The cortico-diaphyseal indices of the tibia were normal (>0.5). |
Mean 25-(OH)D levels dropped from 23 ± 6 ng/mL (1983) to 5.5 ± 2 (1987), (p < 0.03). | No change on radiologic signs of osteopenia during the 4 years of the study. | Not reported |
| Leonberg BL, 1998 [18] | OSCS. Children with IF, weaned off PN in the previous 7 years (at least since 6 months) evaluated for a nutritional and bone disease screening. USA. |
PN for more than 4 months, beginning in the 1st postnatal week, including 1 month of HPN. | BMC of the radius measured by single-photon absorptiometry, compared with age appropriate norms. | 9 children, age 2.8–6 years (mean, SD: 4.9 ± 1), 5 males. Duration of PN: 5–39 months (mean, SD: 14.6 ± 11.4). Duration of HPN: 8.5 ± 5 months. Mean time of PN cessation: 3.4 ±1.4 years | 18 healthy controls matched for age, gender, and race. | Total BMC of SBS children was reduced compared to controls but not different after correction for weight and height (p = 0.8). 44.4% had BMC below normal. |
Median 25-(OH)D levels: 26 ng/mL. | Not evaluated | 75% subjects with low BMC were consuming diets deficient in calcium. No correlation of BMC with PN duration or time since discontinuation. |
| Dellert SF, 1998 [19] | OSCS, case-control. Children with SBS, totally weaned from PN and with a bone disease screening (1984–1994). USA. |
SBS (intestinal resection, omphalocele, gastroschisis) receiving at least 1 month of PN. | BMC measured by DXA in subjects and controls. Serum Ca and P levels were measured. | 18 children, age 2–8 years, 9 males. Duration of PN: 1–67 months (median 7 months). Discontinuation of PN: 1–8 years (median 3 years). | 36 healthy controls matched for age and gender. | BMC of cases was reduced compared to controls. After adjustment for weight and height, there was no difference in BMC between SBS and controls. Serum Ca e P were comparable. |
Mean serum 25-(OH)D were lower in children with SBS respect to controls (26.0 ± 9.8 ng/mL vs. 41.3 ± 8.4 ng/mL, p = 0.002); mean serum 1,25-(OH)2D were similar. |
Not evaluated | BMC was not correlated with PN duration or time since PN discontinuation. |
| Diamanti A, 2010 [20] | OSCS, case-control. Children with IF on a BMD evaluation program (September 2005-September 2007). Italy. |
IF: condition needing PN providing at least 75% of total calories for >4 weeks or at least 50% calories for >3 months. | BMD (BMC, areal BMD, BMAD, BMD z-score) measured by DXA at baseline in patients and controls and after 1 year (in 9 patients). BMC reduction: BMD z-score <−1. Development of fractures was assessed. |
24 children, age 3.5–17.5 years (6.7 ± 5.2 years), 14 males. 7 off PN. Duration of HPN: 3–181 months (median 38). | 24 healthy controls matched for age and gender. Mean age: 6.5 ± 3.9 years, 18 males. | 83% had BMD z-score ≤ −1. BMD z-score, BMD and BMAD were significantly lower in patients compared to controls (p < 0.05 and p < 0.01). Subjects with BMD z-score >−1 were lower in patients (17%) compared to controls (50%). 17% developed fractures. |
Not evaluated. | All DXA variables were increased at the 2nd DXA in 7/9 patients at 1 year follow-up. | No correlation between bone mineral status and PN duration and nutrient intake. Significant correlation between BMC and BMD and weight and height. BMD z-scores were higher in SBS-IF compared to medical causes of IF. |
| Olieman JF, 2012 [21] | OSCS, cross-sectional. Children with infantile SBS (within 1 year) and off PN invited for a nutritional/bone disease screening (January 1975–January 2003). Netherland. |
IF: 70% of resection of SB or PN needed for >42 days after resection or residual SB length <50 cm for preterm and <75 cm for term neonates | BMD LS, BMD TB, BMC, LBM, and %BF were measured by DXA and compared to Dutch reference data. | 40 subjects, 16 males, mean age 14.8 ± 6.8 years, including 31 children, mean age 11.8 ± 4.2 years. Duration of PN: median 110 days (43–2345) |
Dutch reference population. | In children, mean BMC, LBM*, and BMD LS were lower than the reference values (p < 0.05). | Not evaluated. | Not evaluated | Not reported |
| Ubesie AC, 2013 [22] | OSCS, retrospective. IF patients aged ≥3 years under nutritional follow-up in a 5-year period (2007–2012). USA. |
IF: need for PN support for more than 30 days. | Lumbar spine BMD (L1–L4) was measured by DXA (within 6 months from serum evaluation) and compared to reference values. BMD z-score were adjusted for age and height for “short” patients (height <5th percentile on CDC charts). Reduced BMD: BMD ≤ −2 z-score. |
123 IF patients, median age: 4 years (3–22 years), 71 (57.7) % males. PN duration not indicated. 80 had a DEXA scan. |
Reference population (not specified). | 12.5% had a low BMD z-score (adjusted for height). | 40% had vitamin D deficiency (25-(OH)D < 20 ng/mL). | Not evaluated | PN dependence at bone health evaluation was associated with reduced BMD. Increased age was significantly associated with vitamin D deficiency (p = 0.04) and reduced BMD z-score (p = 0.01) at a binary logistic regression analysis. At a logistic regression model, age >10 years was associated only to reduced BMD (p = 0.02). |
| Mutanen A, 2013 [23] | OSCS, cross sectional. Children with IF invited to a bone disease screening (January 1984–August 2010). Finland. |
IF: 50% resection of SB or duration of PN >30 days. | BMD in the left proximal femur and lumbar spine was measured by DXA, and compared with reference values. BMD z-score ≤ −1 was considered low. BMD z-scores were corrected for bone age. Vertebral compression >20% was considered abnormal. PTH levels were measured. Occurrence of fractures was assessed. |
41 IF children, 27 males, mean age 9.9 years (0.2–27). PN duration 41 months (0.7–250). 30 off PN, time after weaning off: 9 years (0.3–27). 30 had a DXA study. |
Reference population (not specified). | BMD z-score was ≤−1 in 70% and ≤−2 in 43%. 3.8% had sustained fractures. 3.8% had vertebral compression. |
41% had vitamin D deficiency (25-(OH)D < 15 ng/mL), and 44% had secondary hyperparathyroidism. Mean 25-(OH)D levels: 21.6 ng/mL on PN; 23.6 ng/mL in patients off PN (p = 0.642). | Not evaluated | At the multiple regression analysis duration of EN after weaning PN, duration of PN and calcium supplementation were the significant predictors for a lower lumbar spine BMD z-score. |
| Appleman SS, 2013 [24] | Observational, case-control, double-center study. PN-dependent IF children invited to a bone disease screening. USA. |
PN-dependent IF: inability to sustain growth without PN | BMC and BMD of the lumbar spine were measured by DXA. PTH was measured. | 20 IF children, median age 26 months (6–127), 9 males (45%). Median time on PN: 18.5 months (4–103) |
49 healthy controls median age 25 months (7–127), 22 males (45%). | Lumbar spine BMC was 15% lower (p = 0.0009), and BMD was 12% lower (p = 0.0035) in IF subjects compared to controls. No differences after adjusting for weight and height. | IF participants had higher serum 25(OH)D than controls (mean levels: 40 vs. 30 ng/mL, p = 0.0005). IF patients had lower prevalence (5% vs. 14%) of vitamin D insufficiency (25(OH)D < 20 ng/mL). |
Not evaluated | No association between aluminum concentration and BMC or BMD. |
| Pichler J, 2014 [35] | OSCS. Children aged >5 years with IF and being on HPN (2002–2010). UK. |
IF: HPN for at least 6 months. | BMD and BMAD were calculated by DXA + bone age assessed by X-ray. Low BMD or BMAD: ≤−2 SDS. Biochemical markers (also ALP) were measured. Occurrence of fractures was recorded. |
45 subjects, 24 males, age 7.7 years (5–18). Time of HPN: 5 years (3.2–12). 18 (40%) on TPN, 26 (58%) on PPN, 1 (2%) off PN. |
UK reference data. | 42% had BMD -≤−2 SDS and 31% had BMAD ≤−2 SDS. Mean BMD SDS was −1.7 ± 1.6; mean BMAD SDS was −1.4 ± 1.5. 37% had a history of fractures (median 1.3), all non-pathological. |
Mean 25(OH)D levels: 57.6 ± 6.44 ng/mL. 7% patients had vitamin D insufficiency (25(OH)D < 25 ng/mL). |
35/42 had repeated DXA at 1 and 2 years. BMD declined significantly at 1 and 2 years, while BMAD only at 1 year. In a multivariate model, the only significant predictor of a change in the BMD SDS was a change in the weight SDS. | No difference of BMD and BMAD SDS according to diagnosis, degree of dependence on PN, and presence of 25-(OH)D deficiency. In a multivariate model, age at the time of DXA was the only predictor of BMD SDS. |
| Derepas C, 2015 [26] | OSCS, case-control. IF children on HPN (June 2012–August 2012). Canada. |
IF: need of PN providing at least 25% calories for more than 6 weeks. | BMD measured annually by DXA on the lumbar spine (L1–L4) and compared to reference values. Reduced BMD: BMD z-score ≤ −1. Serum OC, CTx, BSAP, and PTH levels were assessed. |
13 IF children, age 1.2–10.7 years, 12 males. Duration of PN: 0.5–12.5 years. 9 subjects had DXA (7 results available) |
20 healthy controls matched for age and gender. | Mean serum OC and CTx concentration were lower in IF compared to controls (p < 0.01 and p < 0.05). No differences in BSAP and PTH levels. BMD measured in 9/13 IF subjects ranged from 0.401 to 0.838 g/cm2. BMD z-score ranged from −3.526 to +1.5. 57% subjects had BMD z-score ≤ −1. 28.5% had BMD z-score ≤ −2. |
Not evaluated | Not evaluated | A significant inverse relation was found between BMD and serum OC but no with length of PN, age, weight, or height z-score. BMD z scores was inversely related to CTx also after controling for weight z-scores. |
| Demehri FR, 2015 [27] | OSCS, retrospective. Patients with IF (2006–2012) who underwent DXA screening (at 6 years). USA |
IF: PN dependence of at least 60 days. | BMD measured by DXA at the lumbar spine (L1–L4). BMD z-scores were derived from reference values. MBD was defined as BMD z-score ≤ −1. Serum Ca, P, and PTH measured within 4 weeks of DXA scan. History of pathologic fractures and bone pain. |
36 IF subjects, 21 males 25 off PN (69.4%). Mean PN duration 5.1 ± 5.4 years. 17 subjects had a repeated DXA. |
Reference U.S. population. | Mean BMD z-score was −1.16 ± 1.32. 50% had MBD. 11.1% had pathologic fractures, and 16.6% had history of bone pain. |
Mean 25(OH)D levels: at 1st DXA, 25.11 ± 13.05 ng/mL; at 2nd DXA, 23.67 ± 12.73 ng/mL (p = 0.821). 63.8% had vitamin D deficiency (25(OH)D ≤30 ng/mL), and 25% had hyperparathyroidism (>55 pg/mL). No change in serum 25-(OH)D (p = 0.821), despite an increase in % of subjects on vitamin D supplementation. |
In 17/36 with repeated DXA, no change at the 2nd DXA (after 2 ± 1.1 years) occurred (p = 0.199). | Duration of PN and serum 25-(OH)D were predictors of BMD z-score at univariate analysis. In a multivariate analysis, the only significant predictor of a reduced BMD z-score was the length of PN (B = −0.132, p = 0.006). |
| Khan FA, 2015 [28] | OSCS, retrospective. Patients with IF who underwent DXA screening at 5 years of age (2004–2013). USA. |
Not reported. | Whole-body DXA was performed, and BMD z-scores were determined using normative data. For patients with repeated DXA scans, the lowest BMD z-scores were recorded. MBD: BMD z-score ≤ −2. Serum levels Ca, P, and PTH were measured. History of fractures was recorded. |
65 IF subjects, 34 males. 39 off PN (60%). Mean PN duration 44.2 ± 43.2 months. |
Reference U.S. population. | BMD z-score ≤ −2 was reported in 34%. 29% experienced at least 1 fracture. |
Mean 25(OH)D levels: 27 ± 42 ng/mL. 42% had vitamin D deficiency (25(OH)D ≤30 ng/mL). |
Not evaluated. | At univariate analysis, patients with BMD z-score ≤ −2 had lower WAZ (p = 0.01), lower Ca (p = 0.04), and higher PTH levels (p = 0.006). At the multivariable logistic regression analysis, WAZ (R = 1.8, p = 0.03) and serum Ca (R = 3.8, p = 0.02) were independent factors of a low BMD z-score. |
| Wozniak LJ, 2015 [29] | OSCS, retrospective. IF children under follow-up (January 2012–June 2012) who underwent at least 1 determination of serum 25-(OH)D between July 2010 and June 2012. USA. |
IF: HPN required for at least 6 months. | Osteopenia evaluated on subjective analysis of standard radiology studies (bone mineralization of axial skeleton, long bones, and blurring of the cortical white line). History of fractures was assessed. | 27 IF children, median age 5.5 (IQR: 2.7–8.2 years), 12 (44%) males. Median PN duration: 3.5 years (IQR: 2.6–6.9). | No | 59% had osteopenia. 11% had fractures. |
41% had vitamin D insufficiency (25(OH)D = 20–29 ng/mL); 1 child had vitamin D deficiency (25(OH)D < 20 ng/mL). Supplementation with vitamin D resulted in improvement of 25-(OH)D levels (2/27) and in improvement at blood drawn following the study period (4/27). | Not evaluated. | At univariate and multivariate logistic regression analysis, the only variable associated with 25-(OH)D levels was diagnosis of SBS. Duration of PN was not correlated. |
| Neelis E, 2018 [30] | OSCS, retrospective. IF children (followed between 2000 and 2015) who underwent at least one DXA or DXR. The Netherlands. |
IF: PN needed for >6 weeks. | Total-body and lumbar spine (L2, L2–4) BMD was measured by DXA after 4–5 years of age. BMD z-scores were determined by national reference values. BMD was adjusted to the bone size calculating the BMAD. Low BMD or BMAD: z-score ≤ −2. DXR was used to calculate the BHI that was compared to a reference population. History of fractures was recorded. |
46 IF subjects, 20 (44%) males. Age at 1st DXA: 6 years (IQR: 5.5–9.9). 28 off PN (76%). Median PN duration: 9.4 months (IQR: 4.6–14.3). 37 had 1st DXA. 13 children had multiple DXA scans after a median time of 2.1 years (1.09–2.44). |
Reference Dutch population. | 24.3% had a low BMD (either BMD TB, LS or BMAD z-score ≤ −2) at 1st DXA. Median BMD TB, BMD LS, and BMAD z-scores were significantly lower compared to the reference population (p = 0.006; p < 0.001 and p = 0.004). 50% had low BHI (z score ≤ −2) at 1st radiography. 4/46 children developed multiple fractures (100% were PN dependent at time of 1st fracture). |
33% had vitamin D insufficiency (25(OH)D < 20 ng/mL); no differences between children on PN or weaned off. | Median change in z-scores per year was +0.16 SD for BMD TB and +0.09 for BMD LS. | Patients still on PN at 1st DXA had lower median BMD TB z-score (p = 0.048), and the proportion of children with BMD TB z-scores ≤ −2 was higher (p = 0.008). At univariate analysis, an older age, a lower HFA z-score, a HFA z-score ≤ −2, a higher WFH z-score, and a longer PN duration were related to lower BMD LS z-scores. In a multivariate model older age, longer duration of PN and surgical IF were related to lower z-scores. |
| Poinsot P, 2018 [31] | OSCS, retrospective. IF children (January 2004–January 2014) on HPN with at least 2 DXA (on HPN since at least 2 years from last DXA). France. |
IF: children on HPN for at least 6 months. | TBMC, LTM, and FM were measured by DXA and adjusted for ideal WFH. LBM** was defined as a TBMC z-score ≤ −2, following the reference values. Serum Ca, P, and ALP were measured within 3 months of DXA assessment. |
31 children, 14 males (45%). Median age at 1st DXA: median age: 2.9 years (0.4–13.3). Median PN duration: 2.7 years (0.1–12.7). Median age at last DXA 11.4 years (3.7–19.7), and median PN duration: 9.2 years (2.9–19.6). Median time between 1st and last DXA: 6.2 years (0.7–16.6). |
Reference population composed of 68 healthy children (2–24.9 years, 31 males) and 55 newborns (gestational age 33–40 weeks). | 45% had LBM** at 1st DXA. Median TBMC z-score at 1st DXA was −1.9 SD (−5.3–2.6). 71% had TBMC z-score for ideal WFH ≤−1. | Median 25(OH)D levels: at 1st DXA, 15.4 ng/mL; at last DXA, 14.0 ng/mL. | TBMC z-score adjusted for ideal WFH increased significantly at last DXA (+0.1 ± 0.04 per year, p = 0.012). At last DXA, 29% had LBM (p = 0.197 with 1st DXA). |
IF etiology (CIPO, HD, and CE) and a lower plasmatic creatinine level were related to LBM** prevalence at baseline. At last DXA, LBM** was associated with a shorter PN duration (p = 0.045) and a younger age (p = 0.023). The risk of the LBM** decreased significantly with the duration of HPN (OR 0.9 per year of PN, p = 0.018). |
| Olszewska K, 2018 [32] | OSCS. Patients with ultra-short bowel syndrome on HPN under nutritional follow-up. Poland. |
IF: residual small bowel <10 cm at the time of resection (first 2 months of life) and receiving PN for at least 6 months. | Antero-posterior spinal and total bone mass density measured by DXA, and BMD expressed as z- scores. Decreased BMD TB: ≤−1 z-score. Abnormal bone mineralization: BMDs ≤−1 z-score. |
17 children, age 0.8–14.2 years (median 6.6), 9 males on HPN. Median PN duration: 6.6 years (0.8–14.2 years). 8 subjects had DXA. |
Reference population (not specified). | 50% had BMD ts ≤ −1 z-score, and 87.5% had BMD s ≤ −1 z-score. | Mean 25(OH)D levels: 20.1 ng/mL. 53% had vitamin D deficiency (25(OH)D < 20 ng/mL). |
Not evaluated. | BMD z-scores were not correlated with body mass or body height SDS, age or PN length. |
| Kvammen JA, 2020 [33] | OSCS, cross sectional, controlled. Children with IF on HPN (March 2017–September 2017) who underwent a bone disease screening. Norway. |
IF: children dependent on HPN for >6 months. | BMD was measured by DXA total body (TB) and spinal (LS at L2–L4). BMD z-scores were calculated according to reference values and corrected for bone age (if height was <−2 z-scores). Suboptimal BMD: BMD ≤−1 to −1.99 z-score. Low BMD: ≤−2 z-score. |
19 IF children, mean age: 10.1 years, SD: 3.5, 13 males (68%). 100% on HPN, 4 on TPN. 12 IF children had DXA. |
Reference population composed of 50 healthy controls, mean age: 10 years, SD: 3.6, 18 males (36%). 46 controls had DXA. | IF group had significantly lower median BMD z-score for total body (p < 0.001) and lumbar spine (p = 0.01). 25% IF subjects had suboptimal BMD TB, 17% had suboptimal BMD LS (vs. 5% of healthy controls), and 25% of IF children had low BMD. |
IF group had lower 1,25-(OH)2D levels compared to controls (29.6 vs. 57.5 pg/mL, p < 0.001). IF patients and healthy had similar level of 25(OH)D (28.4 vs. 32.4 ng/mL, p = 0.29). | Not evaluated | Not reported |
| Nader EA, 2021 [34] | OSCS, retrospective. IF children >5 years on HPN, who underwent a DXA screening (January 2016–December 2018). France. |
IF: HPN for at least 2 years | BMD of the lumbar spine (L1–L4), left femur, and total body were measured using DXA and expressed as z-scores (compared to reference values). BMD values were adjusted to height or statural age. Low BMD: ≤−2 z-score. |
40 IF children, 24 males, median age 12.4 ± 4.5 years. HPN duration 12.4 years ± 4.4. | Reference population (not specified). | BMD LS ≤−2 z-score: 30% BMD LS adjusted ≤−2 z-score: 18% LF BMD ≤−2 z-score: 15% LF BMD adjusted ≤−2 z-score: 15% BMD TB ≤−2 z-score: 23% BMD TB adjusted ≤−2 z-score: 18% |
Mean 25(OH)D levels: 26.5 ± 9.1 ng/mL; Median 25(OH)D levels: 25 ng/mL. | Not evaluated | No correlation between indication of PN, duration, and degree of PN dependency and level of 25-(OH)D3. |
| Louazon T, 2021 [36] | Cross-sectional, case-control, single-center study. IF children, aged >9 years, on HPN who underwent a bone assessment (March 2014–June 2015). France |
IF: HPN for at least 2 years | vBMD was evaluated by HR-pQCT at the non dominant limb. BMC and BMD were measured by TB and spine (L1–L4) DXA. Serum NMID osteocalcin and PTH were measured. |
11 IF children, 8 males, median age: 16 years (9–19). Median PN duration: 10.3 years (6.4–18.3). 2 off PN. |
20 healthy controls, 16 males, median age: 16 years (10–17). | In IF children, increased PTH (p = 0.003) and reduced OC-diasorin (p = 0.005). At the tibia, IF children had low trabecular area compared to controls (p = 0.003). BMD TB was lower in IF subjects (p = 0.039). 18% of IF patients had BMC TB ≤−2 DS. |
No differences in mean 25-(OH)D3 levels between patients and controls (17.6 vs. 22.8 ng/mL, p = NS). | Not evaluated | Not reported |
(H)TPN, (home) total parenteral nutrition; (H)PN, (home) parenteral nutrition; IF, intestinal failure; Ca, calcium; P, phosphorus; ALP, alkaline phosphatase; PTH, parathyroid hormone; 25-(OH)D, 25-hydroxyvitamin D; (T)BMC, (total) bone mineral content; SB, short bowel; SBS, short bowel syndrome; BMD, bone mineral density; BMAD, bone mineral apparent density; DXA, dual-energy X-ray absorptiometry; BMD LS, bone mineral density of the lumbar spine; LBM*, lean body mass; BMD SDS, age and sex-adjusted bone mineral density; LTM, lean tissue mass; EN, enteral nutrition; CDC, Center for Disease Control; OC, osteocalcin; CTx, c-telopeptide; BSAP, bone-specific alkaline phosphatase; WAZ, weight-for-age z-score; MBD, metabolic bone disease; BMD TB, bone mineral density total body; HFA, height for age; BHI, Bone Health Index; WFH, weight for height; LBM**, low bone mass; LF BMD, left femur bone mineral density; HR-Pqct, high-resolution peripheral quantitative computed tomography; NMNID, N-terminal mid fragment.