Of the 602 participants, 38 were excluded from this study, such that 566 patients were enrolled. Three hundred two patients were lost to follow-up due to abandoning treatment or returning to a local hospital. Two hundred forty-six patients with newly diagnosed MM were followed for > 6 months. The specific details of enrollment and exclusion are shown in Fig. 1. The baseline characteristics of the study population are listed in Table 1.
Table 1
Demographics of patients with MM
| MM patients in Cross-section analysis group | MM patients followed for at least 6 months |
N | 566 | 246 |
Age(year) | 63.1 ± 10.1 | 61.4 ± 9.7 |
Gender | | |
Female | 220 (38.9%) | 93 (37.80%) |
Male | 346 (61.1%) | 153 (62.20%) |
Hemoglobin(g/l) | 91.2 ± 25.5 | 95.6 ± 24.6 |
Serum globulin(g/l) | 49.8 ± 26.0 | 50.6 ± 26.3 |
Serum albumin( g/l) | 35.1 ± 7.9 | 36.2 ± 8.2 |
serum creatinine level(umol/L) | 184.5 ± 227.9 | 140.0 ± 131.0 |
serum calcium( mmol/l) | 2.33 ± 0.4 | 2.34 ± 0.36 |
albumin-adjusted serum calcium ( mmol/l) | 2.45 ± 0.4 | 2.44 ± 0.38 |
serum kappa light chain( mg/dl) | 2245.7 ± 3098.0 | 2256.96 ± 3048.79 |
serum lamda light chain( mg/dl) | 1380.6 ± 2296.5 | 1274.93 ± 1912.93 |
Urinary kappa light chain( mg/dl) | 187.5 ± 544.3 | 197.86 ± 635.22 |
Urinary lamda light chain( mg/dl) | 324.8 ± 1025.7 | 328.88 ± 1102.17 |
Serum beta 2 microglobulin( ug/l) | 7987.6 ± 7517.0 | 7159.32 ± 6780.31 |
Lactate dehydrogenase( u/l) | 204.5 ± 136.2 | 190.89 ± 98.01 |
Follow-up time.(months) | ---- | 26.64 ± 17.59. |
renal impairment | | |
No | 404 (71.4%) | 188 (76.42%) |
Yes | 162 (28.6%) | 58 (23.58%) |
Antimyeloma Therapy regimen | | |
Traditional chemotherapy | ----- | 159 (65.43%) |
proteasome inhibitor treatment | ----- | 72 (31.17%) |
ESRD | | |
No | --- | 215 (88.48%) |
YES | --- | 28 (11.52%) |
All-cause death | | |
0 | --- | 225 (91.46%) |
1 | ---- | 21 (8.54%) |
Relationship between the serum calcium level and the occurrence of RI based on cross-section analysis
The mean age was 63.1 ± 10.1 years; 61.6% of the 566 patients were men with newly diagnosed MM. Using the IMWG criteria, 162 of the 566 patients (28.6%) with newly diagnosed MM presented with RI.
The median serum calcium level was 2.45 mmol/L. Hypercalcemia was noted in 11.3% of newly diagnosed MM patients when the serum calcium level were was albumin-adjusted, but the incidence of hypercalcemia was 7.42% of newly diagnosed MM patients if the serum calcium level was not albumin-adjusted. The incidence of hypercalcemia in this study was lower than the 20–40% reported in the literature (26,27).
Correlation analysis revealed that the serum calcium level was positively correlated with RI (β = 0.156, p < 0.001), the serum globulin level (β = 0.003, p < 0.001), the serum kappa light chain level (β = 0.001, p = 0.01578), the serum beta-2 microglobulin level (β = 0.001, p < 0.001),and ISS(international staging system) stage (β = 0.230, p < 0.001), and negatively correlated with the eGFR (β = -0.002, p < 0.001), hemoglobin level༈HB༉(β = -0.004, p < 0.001), and albumin level (β = -0.016, p < 0.001).
Based on multivariable linear regression analysis adjusted for these variables, the serum calcium level was independently associated with RI (Table 2).
Table 2
Cross-sectional Correlation Analyses between serum calcium level and the presence of RI in different models
Variable | The occurrence of RI |
Crude model (β, 95%CI, P) | Minimally adjusted model(β, 95%CI, P) | Fully adjusted model (β, 95%CI, P) |
Serum calcium levle(mmol/l) | 3.0 (1.9, 5.0) < 0.001 | 3.0 (1.8, 4.9) < 0.001 | 3.6 (1.8, 7.1) < 0.001 |
Mean serum calcium(mmol/l) | | | |
<2.45 | 1.0 | 1.0 | 1.0 |
>=2.45 | 1.7 (1.2, 2.5) 0.004 | 1.7 (1.2, 2.5) 0.005 | 1.8 (1.1, 3.1) 0.016 |
Note: |
Non-adjusted model adjust for: None |
Adjust I model adjust for: AGE; Gender.; hypertension history; diabetes history |
Adjust II model adjust for: AGE; Gender.; hypertension history; diabetes history, LDH, HB, serum globulin and ISS-stage, serum albumin, serum /Urinary kappa light chain, serum /Urinary lamda light chain and serum beta 2 microglobulin |
Figure 2 is a smoothing plot of the serum calcium level versus RI and eGFR. The curve shows that there was a negative correlation between the serum calcium level and the eGFR, and the relationship between the serum calcium level and RI was not simply linear. Specifically, as shown in Table 3, threshold effect analysis indicated that the incidence of RI increased as the serum calcium level increased, up to 2.30 mmol/L. In patients with a serum calcium level < 2.30 mmol/L, the serum calcium level was not significantly associated with RI in MM patients (β = 0.10, p = 0.108). In patients with a serum calcium level > 2.30 mmol/L, the correlation coefficient of RI was positive (β = 6.2, p < 0.001).
Table 3
Threshold Effect Analysis of serum calcium level and the presence of RI using Piece-wise Linear Regression
Model | Result [β (95%CI) P value] |
Model I | |
one-line linear regression model | 2.80 (1.4, 5.6) 0.004 |
Model II | |
turning point | 2.30 mmol/L |
Group1 < 2.3 correlation coefficient (β1) | 0.10 (0.0, 2.3) 0.108 |
Group2 > 2.3, correlation coefficient (β1) | 6.20(1.9, 10.2) <0.001 |
predictive value of RI at turning point | -1.40 (-1.7, -1.1) |
a log likelihood ratio test | 0.011 |
Note: |
Effect: albumin-adjusted serum calcium level Cause: Myeloma-Related RI |
Adjusted for AGE; Gender; hypertension history; diabetes history LDH, serum /Urinary kappa light chain, serum /Urinary lamda light chain,HB, Serum globulin and ISS-stage, serum albumin ,serum beta 2 microglobulin |
Serum calcium level, ESRD, and all-cause death
To evaluate the prognostic value of the serum calcium level in MM patients followed for > 180 days, patients were categorized into 2 groups according to the mean serum calcium level(2.44 mmol/L), and the clinical outcomes were compared (Table 4).
Table 4
Clinical Outcomes according to mean serum calcium level
Variable | mean serum calcium level < 2.44umol/L | mean serum calcium level > 2.44umol/L | P-value* |
N | 160 | 86 | |
Age(year) | 60.7 ± 9.8 | 62.7 ± 9.6 | 0.197 |
Follow-up time(month) | 27.4 ± 18.8 | 25.2 ± 15.1 | 0.730 |
Gender. | | | 0.23 |
female | 62 (38.8%) | 31 (36.0%) | |
Male | 98 (61.3%) | 55 (64.0%) | |
ESRD. | | | 0.970 |
No | 139 (88.5%) | 76 (88.4%) | |
Yes | 18 (11.5%) | 10 (11.6%) | |
All-cause death. | | | 0.015 |
No | 152 (95.0%) | 73 (84.9%) | |
Yes | 8 (5.0%) | 13 (15.1%) | |
Patients who were followed for at least 6 months were included in the study. A total of 246 patients were studied in this observational cohort. The mean age was 61.41 ± 9.73 years and 62.20% were men. The mean duration of follow-up was 26.64 months. Twenty-one patients (8.54%) died and 28 patients(11.52%)had ESRD after diagnosis. Of these patients, 18 (11.46%) with a mean serum calcium level < 2.44 mmol/L developed ESRD compared with 13 patients (11.6%) who had a serum calcium level > 2.44 mmol/L (p > 0.05). Moreover, all-cause deaths occurred in 13 patients (15.1%) with a mean serum calcium level > 2.44 mmol/L and 8 patients (5.00%) with a mean serum calcium level < 2.44 mmol/L (p < 0.05; Table 4).
Correlation analyses based on univariate regression revealed that all-cause deaths correlated with the serum calcium level (β = 4.817, p < 0.001), Lactate dehydrogenase (LDH)(β = 1.004, p < 0.001), RI (β = 4.817, p = 0.05) and eGFR (β = 0.983, p = 0.01).
In Cox regression analysis adjusted for demographic and clinical factors of age, gender, an increase in the serum calcium level was significantly associated with an increase in the risk of the all-cause death (HR, 4.82; 95% CI, 2.37–9.78; p < 0.001; Table 5 [model 1]). Furthermore, the fully adjusted model, including RI, LDH, HB, albumin, serum beta 2 microglobulin, serum /urinary kappa or lamda light chains, and anti-myeloma therapy regimen, showed a significant increase in risk of adverse outcomes conferred by the serum calcium level (HR, 5.72; 95% CI, 2.09–15.63; p < 0.001; Table 5 [model 2]). We also did a separate analysis in which the serum calcium level was treated as a categorical variable by the mean. The risk of all-cause death was significantly higher in patients with a mean serum calcium level > 2.44 mmol/L than in patients with a serum calcium level < 2.44 mmol/L (HR, 6.99; 95% CI, 1.61–30.41; p < 0.01; Table 5). A Kaplan-Meier curve revealed that event-free survival for the all-cause death outcome was significantly lower in patients with a serum calcium level > 2.44 mmol/L compared to patients with a serum calcium level < 2.44 mmol/L (p = 0.0063; Fig. 3).
Table 5
༎Relationship between serum calcium level and All-cause death and ESRD in different models
Variable | All-cause death | ESRD |
Crude model (β, 95%CI, P) | Minimally adjusted model (β, 95%CI, P) | Fully adjusted model (β,95%CI,P) | Crude model (β, 95%CI, P) | Minimally adjusted model (β, 95%CI, P) | Fully adjusted model (β, 95%CI, P) |
Serum calcium levle(mmol/l) | 4.82 (2.37, 9.78) < 0.0001 | 4.95 (2.36, 10.37) < 0.0001 | 5.72 (2.09, 15.63) 0.0007 | 1.34 (0.53, 3.37) 0.5331 | 1.32 (0.51, 3.40) 0.5624 | 0.85 (0.29, 2.48) 0.7655 |
Mean serum calcium(mmol/l) | | | | | | |
<2.44 | 0 | 0 | 0 | 1.0 | 1.0 | 1.0 |
>=2.44 | 4.13 (1.62, 10.51) 0.0029 | 4.27 (1.66, 11.01) 0.0026 | 6.99 (1.61, 30.41) 0.0095 | 1.09 (0.50, 2.39) 0.8211 | 1.03 (0.47, 2.25) 0.9444 | 0.95 (0.39, 2.34) 0.9139 |
Note: |
Non-adjusted model adjust for: None |
Adjust I model adjust for: AGE; Gender.; hypertension history; diabetes history |
Adjust II model adjust for: AGE; Gender.; hypertension history; diabetes history,LDH, serum /Urinary kappa light chain, serum /Urinary lamda light chain, HB, serum globulin and ISS-stage, serum albumin and serum beta 2 microglobulin |
In this study we showed that the serum calcium level does not predict ESRD in a Cox regression model (HR = 0.85; 95% CI, 0.29, 2.48; p = 0.77 l; Table 5). An increase in the serum calcium level was not associated with an increased risk of ESRD(Fig. 4).
To evaluate the operating characteristics of the serum calcium level as a prognostic value for all-cause death in MM patients followed for > 6 months, we conducted a time-dependent ROC analysis for the serum calcium level in comparison with RI (Fig. 5). The areas under the ROC curve for the serum calcium level at 12, 27, and 46 months were 0.705, 0.699, and 0.763, respectively; however, the areas under the ROC curve for RI were 0.736, 0.63, and 0.573, respectively (p = 0.225).