A total of 105 ITP patients were recruited from the First People's Hospital of Huzhou (Huzhou, China) and Huzhou Central Hospital (Huzhou, China) between January 2014 and June 2017. The patients were between 18 and 50 years of age and provided written informed consent to participate in the study. The inclusion and exclusion criteria are presented in Fig. 1. Among the candidate ITP patients, 7 post-menopausal women were excluded due to their clinical presentation, pelvioscopy results and serum levels of follicle-stimulating hormone (>10 U/l) and estradiol (<20 pg/ml). The baseline physiological and biochemical parameters of all patients are provided in Tables I and II. The effect of steroids on bone loss in ITP patients mainly depends on the accumulated dose. Before the commencement of the present trial, all ITP patients received steroid therapy for one month. At 1 month prior to the start of the trial, the steroid treatment was discontinued for all ITP patients, observation or replacement therapy with gamma globulin.

Patient randomization was performed using the SAS 9.0 system (SAS Institute, Cary, NC, USA). The 40 enrolled ITP patients were randomized into CalD or CalD + ALN treatment groups at a pre-determined proportion of 1:1. The randomization process was an unsupervised classification. Before the start of the trial, the participants were given a fixed prescription and received study medications from pharmacy staff according to the enrollment sequence. The medication for the CalD and CalD + ALN groups had the same appearance, shape, color and packaging, so that the research physicians and participants were not distinguish them. The research physicians and participants were not informed about any individual treatment details until the end of the treatment course and after receiving the laboratory test results. The Huzhou Central Hospital (Huzhou, China) approved the study protocol (approval no. 201401A006). Informed consent forms were signed by all patients. None of the patients were included in the current study prior to obtaining ethical approval.

The patients in the CalD group (n=20) received 1,200 mg calcium and 250 IU vitamin D₃ daily (Caltrate D; Pfizer, Inc., New York, NY, USA). Patients in the CalD + ALN group (n=20) received 1,200 mg calcium and 250 IU vitamin D₃ daily with ALN (70 mg; Merck & Co., Darmstadt, Germany) weekly for 9 months.

The BMD from the first to fourth lumbar vertebrae (L1-L4), femoral neck and total hip were measured by dual-energy X-ray absorptiometry (Hologic DQR-4500W; Hologic, Inc., Bedford, MA, USA), according to the manufacturer's protocol (16). Osteoporosis and osteopenia were defined according to international guidelines: Osteoporosis was classified by a t-score of <-2.5 standard deviation (SD) in either lumbar spine or femur, and osteopenia by a t-score between −2.5 and −1 SD (18).

Serum and urine samples for determination of Ca and P were obtained at baseline levels and after 1, 3, 6 and 9 months, using an Olympus AU 2700 automated multichannel analyzer (Olympus, Tokyo, Japan). Bone formation and resorption markers in the serum were determined at the baseline and after 1, 3, 6 and 9 months. Serum intact parathyroid hormone (PTH; cat. no. 11972219122) and osteocalcin (cat. no. 12149133122) were measured using chemiluminescence immunoassays (Roche Diagnostics, Mannheim, Germany). Serum 25(OH) vitamin D was determined using a DiaSorin kit on a LIAISON automated immunoassay analyzer (DiaSorin, Saluggia, Italy). Serum bone-specific alkaline phosphatase (cat. no. 48262; BAP; Access Ostase, Beckman Coulter, Inc., Brea, CA, USA), type 1 procollagen N-terminal propeptide (cat. no. E-EL-H0185c; P1NP; Elabscience Biotechnology Co., Ltd, Wuhan, China), tartrate resistant acid phosphatase 5b (cat. no. CSB-E08490h; TRACP-5b; CUSABIO Technology LCC, Wuhan, Hubei, China) and C-terminal telopeptides of type 1 collagen (cat. no. E-EL-H0960c; CTX; Elabscience Biotechnology Co., Ltd, Wuhan, China) were measured with a commercial ELISA. All tests were performed according to the manufacturers' instructions in a routine laboratory.

Continuous data are expressed as the mean ± standard deviation. A two-tailed, unpaired Student's t-test for independent samples or an χ² test was used to compare baseline variables between the two groups. The differences between quantitative variables were analyzed using a Student's t-test for unpaired data. Inter-group differences were analyzed by one-way analysis of variance, followed by Tukey's post-hoc test. SPSS (v.17; SPSS, Inc., Chicago, IL, USA) for Windows was used for statistical analysis and GraphPad Prism software (version 7.0; GraphPad Software, Inc., La Jolla, CA, USA) was used to draw the histograms. P<0.05 was considered to indicate statistically significant differences.

Initially, 105 patients with a clinical diagnosis of ITP and who had received steroids treatment were recruited for the present study. Among these patients, 67 were included based on the following: i) An age of 18–50 years; ii) no use of steroids, calcium, vitamin D, anti-inflammatory agents or anti-osteoporotic drugs within one month prior to the trial; iii) the patients provided written informed consent form. In addition, 27 patients were excluded due to co-morbidities or post-menopausal status, and the remaining 40 patients were randomized into two groups (20 in the ALN + CalD group and 20 in the CalD group; Fig. 1).

The baseline clinicopathological characteristics and biochemical parameters of the patients with ITP are presented in Tables I and II, respectively. At the baseline, clinicopathological characteristics, including the gender ratio, age, BMI, steroid dosage and therapy duration, platelet count, as well as the BMD of lumbar vertebrae, femoral neck and total hip, were not significantly different between the ALN + CalD and CalD groups (P>0.05; Table I). Furthermore, the baseline levels of serum Ca, P, intact PTH, 25(OH) vitamin D, BAP, P1NP, osteocalcin, TRACP-5b, CTX and 24-h urinary Ca were not significantly different between the two groups (Table II).

First, the association between BMD (T-score) and the average daily steroid dose was demonstrated in patients with ITP. The BMD T-score was −0.64±0.28 in the group receiving >5 mg/day, −1.54±0.48 in the group treated with 5–10 mg/day, −2.74±0.30 in the group receiving 10–15 mg/day and −3.34±0.66 in the group receiving >15 mg/day, suggesting that average daily steroid dose was positively correlated with bone loss in ITP patients with steroids treatment prior to clinical trial (Fig. 2).

The BMD of the lumbar vertebrae L1-L4 was significantly increased between month 3 and 9 of ALN + CalD treatment compared with that at the baseline. Compared with the baseline levels, CalD treatment alone also increased the BMD of the lumbar vertebrae L1-L4 at month 9 in patients with ITP. These results suggest that ALN + CalD and CalD have a beneficial effect to improve the BMD of the lumbar vertebrae L1-L4 in patients with ITP. Of note, the BMD of the lumbar vertebrae L1-L4 in the ALN + CalD group was significantly higher than that in the CalD group at month 9 (P<0.05; Table III; Fig. 3A). In addition, the BMD of the femoral neck and total hip was significantly increased in the ALN + CalD group compared with that at the baseline at month 6 and 9. However, CalD treatment alone had no significant effect on the BMD of the femoral neck and total hip in patients with ITP. Of note, ALN combined with CalD markedly increased the BMD of the femoral neck and total hip as compared with that in the CalD group at month 9 (P<0.05; Table III; Fig. 3B and C). These results indicate that ALN combined with CalD is superior to CalD treatment alone in protecting against steroid-associated bone deterioration in patients with ITP.

No change of serum Ca from the baseline was observed at each time-point in each of the two treatment groups (Fig. 4A). There was no difference in serum Ca between the CalD and ALN + CalD groups, except at month 1. Of note, after 9 months of ALN + CalD treatment, the 24-h urinary Ca excretion was significantly suppressed compared with the baseline levels. Compared with CalD treatment alone, ALN + CalD caused a significant decrease in 24-h urinary Ca at 9 months. However, CalD treatment alone did not significantly change the 24-h urinary Ca at any time-point (Fig. 4B). Compared with the baseline levels or CalD treatment alone, ALN + CalD treatment markedly decreased serum P at month 1 and 9. CalD treatment alone had on significant effect on serum P at any time-point (Fig. 4C). Neither CalD treatment alone nor CalD combined with ALN had any effect on PTH at any time-point (Fig. 4D). An increase of 25(OH) vitamin D levels was detected in the CalD and CalD combined with ALN groups after 3 months of treatment; however, no significant difference in 25(OH) vitamin D levels was observed between the two groups (Fig. 4E).

The present study also indicated that CalD treatment alone or combined with ALN provided no significant improvement in the bone formation markers BAP, P1NP and osteocalcin (Fig. 5A, B and C, respectively), while the bone absorption markers TRACP-5b and CTX were inhibited by CalD combined with ALN treatment (Fig. 5D and E). Compared with the baseline levels, ALN + CalD treatment resulted in a significant decrease in TRACP-5b levels after 3 months of treatment. TRACP-5b was significantly lower in the ALN + CalD group compared with that in the CalD group at month 6 and 9 (Fig. 5D). Compared with the baseline levels, ALN + CalD treatment led to a reduction of CTX levels in the serum after 9 months of treatment. The levels of CTX were significantly decreased in the ALN + CalD group compared with those in the CalD group. These results suggested that ALN had a specific effect to inhibit bone resorption markers in patients with ITP.