Risk factors for treatment-related bone loss and osteoporosis in patients with follicular lymphoma

Abstract

The survival rate of non-Hodgkin lymphoma (NHL) has steadily improved. However, osteoporosis introduced by treatment is prevalent and associated with increased mortality and disability for patients with NHL. We aimed to investigate factors impacting bone mineral density (BMD) reduction and osteoporosis, and the trend of BMD after chemotherapy. Overall, 97 newly diagnosed patients with follicular lymphoma (FL) were retrospectively enrolled. CT attenuation values were measured to assess BMD levels. Although 73.2% of patients received calcium and vitamin D supplements, 44.3% showed significant BMD reduction, and baseline BMD and hemoglobin levels were the risk factors. 26.6% of patients newly developed osteoporosis post-chemotherapy where age and cumulative dose of glucocorticoid were risk factors. The results of 20 patients with consecutive follow-up showed that BMD continued to decline for 6 months post-chemotherapy and did not return to baseline values. Therefore, BMD evaluation and more positive anti-resorption treatments should be administered for high-risk patients.

Keywords:

• Follicular lymphoma
• osteoporosis
• CT attenuation value
• bone mineral density
• chemotherapy

Introduction

Follicular lymphoma (FL) is one of the most prevalent non-Hodgkin lymphoma (NHL) subtypes, accounting for 8.1 ∼ 23.5% of all NHL cases in China, with incidence rising year by year [1]. Clinically, first-line treatments including chemotherapy and targeted therapy have been applied to improve the 5-year survival rate of FL. However, R-CHOP (rituximab + cyclophosphamide, doxorubicin, vincristine, prednisone), R-CVP (rituximab + cyclophosphamide, vincristine, prednisone), and other first-line therapy options including glucocorticoids (GCs), cyclophosphamide and doxorubicin, are inevitably associated with several side effects [2, 3], attracting widespread attention to reducing treatment-related problems and improving quality of life. In recent years, several studies have found a link between osteoporosis or fracture and patients following NHL treatment [4, 5]. As one of the major contributors to increased mortality and disability, there is a need for improved prevention of osteoporosis and osteoporosis-related fractures.

The National Comprehensive Cancer Network (NCCN) 2023 guidelines for B-cell lymphomas recommend post-treatment bone mineral density (BMD) evaluation within 1 year, and symptomatic treatment with calcium and vitamin D for deficient patients, while bisphosphonate therapy is recommended only in patients with osteoporotic BMD or osteoporosis-related fractures. However, many patients still experience osteoporosis or even fractures after treatment. Early identification and early intervention are important for such patients. Studies have explored the factors causing BMD reduction and osteoporosis to better predict osteoporosis and preserve bone tissue [4–6]. Besides, the dose-response relationship and the continuous trend of BMD remain unclear, and there is a lack of long-term follow-up and research on single subtypes.

Furthermore, routine computed tomography (CT) scans (including abdominal and thoracic CT) and positron emission tomography-CT (PET-CT) are more frequently used in most patients with NHL for diagnosis. Recently, several studies have found that lumbar vertebral CT attenuation values correlate significantly with dual-energy X-ray absorptiometry (DXA) results [7, 8], which is the gold standard for osteoporosis diagnosis. Therefore, CT attenuation values can be used to measure BMD in patients with NHL [9–11]. When using CT attenuation value to measure BMD, 135 Hounsfield units (HU) at the first lumbar vertebra (L1) is used as the cutoff value to diagnose osteoporosis.

Our study intended to find the correlation between vertebral CT attenuation values with BMD levels and collect relevant clinical parameters of patients with FL in our center; therefore, we investigating the factors causing osteoporosis and BMD reduction, and the continuous trend of BMD after chemotherapy. The long-term goal of our study was to provide further evidence for the prevention and treatment of osteoporosis and related fractures in patients with NHL.

Materials and methods

Patient population

Between January 2014 and December 2020, newly diagnosed patients with FL were analyzed retrospectively. The inclusion criteria were as follows: 1) FL was diagnosed, and 2) an L1-containing CT (including PET-CT, thoracic CT, or abdominal CT) was performed within 3 months of the start and end of chemotherapy. The exclusion criteria were as follows: 1) a previous history of other tumors; 2) history of drugs that affect BMD (bisphosphonates and GCs); 3) history of vertebral compression fracture before chemotherapy; 4) incomplete chemotherapy schedule or no chemotherapy; 5) underwent radiotherapy; 6) progressive disease; and 7) Eastern Cooperative Oncology Group (ECOG) performance status scale > 2.

None of the patients underwent DXA scans. CT attenuation values of L1 were used instead of T-values. In CT scans, L1 is easy to identify, and the image of L1 is included on all thoracic and abdominal CTs, allowing for the most efficient utilization of previous data. The threshold for diagnosis of osteoporosis was set at 135 HU, as suggested by previous studies, with a sensitivity and specificity of approximately 75%. In addition, the findings of previous studies in our center on the correlation between CT attenuation values and DXA T-values were similar to those of the study by Pickhardt et al. [12]. The rate of BMD reduction was calculated by dividing the difference of CT attenuation value by the attenuation value before chemotherapy, and 15% was defined as the cutoff value for significant change.

The clinical parameters collected from patient medical records, including age, sex, body mass index (BMI), FL grade, Ann Arbor Stage, calcium + Vitamin D supplement, chemotherapy protocols, treatment outcome, GCs administration, and other parameters. Based on five parameters (age, bone marrow involvement, hemoglobin [Hb], beta 2 microglobulin [β2 mg], and diameter of the largest affected lymph node), the Follicular Lymphoma International Prognosis Index (FLIPI-2) score was calculated and then divided into low-risk (0–1), medium risk (2), and high risk (≥ 3). All the above parameters comprised the baseline data before chemotherapy. The dosage of GCs was evaluated by cumulative dose (CD), cumulative dose per body surface and daily dose (DD). The CD was calculated as the sum of the equivalent prednisone doses used during treatment, and the DD was calculated as the CD divided by the duration of treatment. The body surface was calculated by the formula: 0.0061 × height + 0.0128 × weight-0.1529.

CT attenuation values measurement

The CT data of the patients were obtained from PET-CT or other common CTs performed during peri-chemotherapy and follow-up. The aim of the examination was to validate the diagnosis and condition of a specific organ while requiring no additional cost, patient time, or radiation exposure. The PET-CT instrument we employed was the Siemens Biograph 64 PET/CT system (Erlangen, Germany). The thoracic or abdominal CT was performed on a Siemens Dual Source CT (Erlangen, Germany) with tube voltages of 120 kV. The interval between the follow-up CT and the previous examination was ≥ 6 months. Furthermore, if a recurrence was found during follow-up, subsequent CT data were discarded because of the potential effect of the disease on BMD.

The CT data were evaluated on a PACS workstation, and the non-contrast scan phase was selected. The measurement method was the same as described previously [13]. The CT attenuation values were assessed on axial L1 images of CT scans (except for one case involving the L1 vertebral body, where the data were derived from the average values for T12 and L2). The measurement plane was selected between the midline zone and the superior endplate [8], with the intervertebral disk and cortical bone excluded from the sagittal images. An ovoid range of interest (ROI) included the anterior two-thirds of the cancellous bone as much as possible while avoiding the venous sinus and focal abnormalities. The HU value was the average attenuation value of the ROI shown in the PACS (Figure 1).

Figure 1. Example of CT attenuation value measurement.

The first non-rib-bearing vertebra was designated as L1. In the sagittal view of CT, the measurement area was chosen between the midline zone and the superior endplate. An ovoid ROI was positioned on the anterior two-thirds of the vertebra, taking care to include the trabecular bone while avoiding the venous sinus and focal abnormalities.

Statistical methods

SPSS Statistics 26.0 software (Ehningen, Germany) was used to analyze the data. The mean and standard deviation were used to describe continuous variables with normal distributions. The median and interquartile range were used to characterize continuous variables without a normal distribution. Categorical variables were expressed as frequency and percentage. Based on different variables (Student’s t-test, paired t-test, Mann–Whitney U test, and Chi-square test), statistical tests were two-sided with a significance level of 5%. The factors with possible statistical significance (p < 0.1) in univariate analysis were included in the logistic regression analysis. The continuous variables were converted to ordinal categorical variables. Each variable was divided into four groups based on its quartiles and assigned a value ranging from 1–4 from small to large. One-way repeated measures (ANOVA) were used to analyze the continuous trend of BMD post-chemotherapy. The Bonferroni method was used for the post hoc test.

Results

Cohort characteristics

There were 151 newly diagnosed patients with FL in our center from January 2014 to December 2020. Of these, 29 had incomplete data. An additional 25 patients were excluded, of whom five had previous tumors (two with breast cancer, one with renal cancer, one with lung adenocarcinoma, and one with endometrial cancer), four had vertebral compression fracture before diagnosis, two had incomplete chemotherapy, six were without chemotherapy, three were undergoing radiotherapy (overlapping with those without or incomplete chemotherapy), four were administered drugs that affected BMD (two with GC and bisphosphonates for rheumatoid disease, one with GCs for myasthenia gravis, and one with GCs and Tacrolimus for kidney transplantation), two showed progressive disease, and two had ECOG performance status scale > 2.

Table 1 shows the characteristics of the 97 patients finally included in this study. At the start of chemotherapy, 86 patients underwent a PET-CT scan, 8 underwent a thoracic CT scan, and 3 underwent an abdominal CT scan. The median interval between the examination and the start of chemotherapy was 5 d (range 1 ∼ 83 d). At the end of chemotherapy, 67 patients underwent PET-CT scans, 20 underwent thoracic CT scans, and 10 underwent abdominal CT scans. The median interval between the first examination and the chemotherapy date was 21 d (range 0 ∼ 79 d). The median interval between pre- and post-chemotherapy examinations was 191 d (range 76 ∼ 490 d). Chemotherapy protocol included RCHOP (60/61.9%) and RCHOP + RFC (rituximab + fludarabine, cyclophosphamide) (20/20.6%). The remaining rare protocol included R (rituximab) (5/5.2%), RCHOP + R (3/3.1%), RCHOP + RFMD (rituximab + fludarabine, mitoxantrone, dexamethasone) (3/3.1%), RCHOP + RB (rituximab + bendamustine) (2/2.1%). Four patients received more than 3 kinds of medication regimens. Before chemotherapy, the CT attenuation value was 169.6 ± 38.8 HU; after chemotherapy, the CT attenuation value was 148.7 ± 42.6 HU. The difference between them was statistically significant (p < 0.001) at 20.9 ± 18.9 HU. The mean proportion of BMD reduction (ΔBMD) after chemotherapy was 12.76 ± 12.13%.

Table 1. Cohort characteristics.

	All (n = 97)
Age	49.85 ± 11.99
Sex (Male)	41(42.3%)
BMI	23.9 (21.1 ∼ 26.0)
Grade
I-II	69(71.1%)
III	28(28.9%)
Ann Arbor Stage
I-II	17(17.5%)
III-IV	80(82.5%)
B symptoms (N/Y)	72(74.2%)/25(25.8%)
Bone marrow involvement (N/Y)	50(51.5%)/47(48.5%)
FLIPI-2
0-1	67(69.1%)
2	15(15.5%)
≥3	15(15.5%)
Smoking history (N/Y)	79(81.4%)/18(18.6%)
Calcium + Vitamin D supplement (N/Y)	26(26.8%)/71(73.2%)
Protocol
R-CHOP	60(61.9%)
R-CHOP + RFC	20(20.6%)
Other	17(21.5%)
Effect
CR	81(83.5%)
PR	11(11.3%)
SD	5(5.2%)
CD (mg)	3000(1975 ∼ 3500)
CD pre body surface (mg/m2)	1592(1013 ∼ 2096)
DD (mg/d)	15.9 ± 8.3
Pre-CT (HU)	169.6 ± 38.8
Post-CT (HU)	148.7 ± 42.6
ΔBMD (%)	12.76 ± 12.13

CR: complete response; PR: partial response; SD: stable disease; CD: cumulative dose; DD: daily dose; Pre-CT: CT attenuation value before chemotherapy; Post-CT: CT attenuation value after chemotherapy; ΔBMD: proportion of BMD reduction.

BMD at diagnosis

There were 18 of 97 patients with a baseline CT attenuation value ≤ 135 HU. The remaining 79 patients comprised the control group. A study was conducted between the two groups. Table 2 shows the results of the univariate analysis. Age, grade, alkaline phosphatase (ALP), albumin (Alb) and creatinine clearance rate (CCr) were statistically significant (p < 0.05), and bone marrow involvement was borderline significant (p < 0.1). The multivariate analysis for the prediction of osteoporosis pre-chemotherapy showed that age was a risk factor (Odds Ratio [OR] = 8.54,95% Confidence Interval [CI] 2.38 ∼ 30.64, p < 0.001) (Figure 2a).

Figure 2. Logistic analysis.

(a) Related factors of BMD osteoporosis before chemotherapy; (b) Related factors of significant BMD reduction; (c) Related factors of osteoporosis after chemotherapy. Hb: hemoglobin; ALP: alkaline phosphatase; Alb: albumin; LDH: lactate dehydrogenase; Ca: calcium; Cr: creatinine; CCr: creatinine clearance rate; β2 mg: beta 2 microglobulin; CD: cumulative dose; Pre-CT: CT attenuation value before chemotherapy.

Table 2. Univariate analysis of BMD at diagnosis.

	Group Cpre(n = 79)	Group Opre (n = 18)	p-value
Age	46.9 ± 10.8	62.9 ± 7.1	<0.001**
Sex			0.748
Male	34(43.0%)	7(38.9%)
Female	45(57.0%)	11(61.1%)
BMI	24.2(20.9 ∼ 26.4)	22.0(21.4 ∼ 24.3)	0.150
Grade			0.028**
I-II	60(75.9%)	9(50.0%)
III	19(24.1%)	9(50.0%)
Ann Arbor Stage			1.000
I-II	14(17.7%)	3(16.7%)
III-IV	60(82.3%)	15(83.3%)
B-symptom			1.000
N	58(73.4%)	14(77.8%)
Y	21(26.6%)	4(22.2%)
Bone marrow involvement			0.052*
N	37(46.8%)	13(72.2%)
Y	42(53.2%)	5(27.8%)
FLIPI-2			0.792
0-1	55(69.6%)	12(66.7%)
2	13(16.5%)	2(11.1%)
≥ 3	11(13.9%)	4(22.2%)
Smoking history			1.000
N	64(81.0%)	15(83.3%)
Y	15(19.0%)	3(16.7%)
Calcium + Vitamin D			0.241
N	19(24.1%)	7(38.9%)
Y	60(75.9%)	11(61.1%)
Effect			0.630
CR	65(82.3%)	16(88.9%)
PR	9(11.4%)	2(11.1%)
SD	5(6.3%)	0(0.0%)
Hb (g/L)	132(120 ∼ 150)	128(121 ∼ 143)	0.259
ALP (U/L)	68(58-85)	81(69-98)	0.035**
Alb (g/L)	43.8(42.1 ∼ 46.2)	40.4(37.6 ∼ 44.7)	0.028**
LDH (U/L)	196(173-228)	212(176-266)	0.341
Ca (mmol/L)	2.30 ± 0.13	2.30 ± 0.13	0.935
Cr (μmol/L)	74 ± 15	69 ± 12	0.154
CCr (ml/min)	95(82-110)	76(66-87)	0.001**
β2-mg (mg/L)	2.15(1.61-2.55)	2.24(1.71-2.67)	0.590

Group Cpre: patients with normal BMD at the time of FL diagnosis; Group Opre: patients with BMD ≤ 135HU at the time of FL diagnosis; CR: complete response; PR: partial response; SD: stable disease; Hb: hemoglobin; ALP: alkaline phosphatase; Alb: albumin; LDH: lactate dehydrogenase; Ca: calcium; Cr: creatinine; CCr: creatinine clearance rate; β2 mg: beta 2 microglobulin.

* p < 0.1.

** p < 0.05.

BMD reduction during treatment

The proportion of BMD reduction was calculated using the ratio of CT attenuation value difference before and after chemotherapy to the baseline value. Forty-three of 97 patients (44.3%) had a BMD reduction >15% during chemotherapy. Table 3 shows the results of the univariate analysis. Hb, CCr, β2 mg, and the baseline CT attenuation value were associated with significant BMD reduction (p < 0.05). B symptoms and FLIPI-2 were borderline significant (p < 0.1). The multivariate analysis for the prediction of significant BMD reduction identified the following parameters as factors affecting BMD reduction: Hb (OR = 0.55, 95%CI 0.32 ∼ 0.94, p = 0.029) and baseline CT attenuation value (OR = 0.60, 95%CI 0.38 ∼ 0.94, p = 0.024) (Figure 2b).

Table 3. Univariate analysis of significant BMD reduction.

	ΔBMD ≤ 15% (n = 54)	ΔBMD > 15% (n = 43)	p-value
Age	48.5 ± 12.5	51.5 ± 11.2	0.230
Sex			0.368
Male	25(46.3%)	16(37.2%)
Female	29(53.7%)	27(62.8%)
BMI	24.2(21.4 ∼ 26.2)	22.6(20.9 ∼ 25.1)	0.307
Grade			0.524
I-II	37(68.5%)	32(74.4%)
III	17(31.5%)	11(25.6%)
Ann Arbor Stage			0.173
I-II	12(22.2%)	5(11.6%)
III-IV	42(77.8%)	38(88.4%)
B-symptom			0.067*
N	44(81.5%)	28(65.1%)
Y	10(18.5%)	15(34.9%)
Bone marrow involvement			0.196
N	31(57.4%)	19(44.2%)
Y	23(42.6%)	24(55.8%)
FLIPI-2			0.053*
0-1	41(75.9%)	26(60.5%)
2	9(16.7%)	6(14.0%)
≥ 3	4(7.4%)	11(25.6%)
Smoking history			0.991
N	44(81.5%)	35(81.4%)
Y	10(18.5%)	8(18.6%)
Calcium + Vitamin D			0.827
N	14(25.9%)	12(27.9%)
Y	40(74.1%)	31(72.1%)
Effect			0.270
CR	43(79.6%)	38(88.4%)
PR	8(14.8%)	3(7.0%)
SD	3(5.6%)	2(4.6%)
CD (mg)	3000(1804 ∼ 3245)	2800(2000 ∼ 4000)	0.408
CD pre body surface (mg/m2)	1606(930 ∼ 2007)	1493(1140 ∼ 2297)	0.389
DD (mg/d)	16.9 ± 9.1	14.8 ± 7.2	0.219
Hb (g/L)	139(125 ∼ 155)	127(113 ∼ 137)	0.001**
ALP (U/L)	68(59–84)	74(61–94)	0.149
Alb (g/L)	43.9(42.3 ∼ 46.4)	42.9(38.3 ∼ 45.4)	0.102
LDH (U/L)	199(174–225)	201(171–251)	0.532
Ca (mmol/L)	2.33 ± 0.13	2.28 ± 0.12	0.056*
Cr (μmol/L)	74 ± 13	63 ± 16	0.765
CCr (ml/min)	94(81–110)	87(76–105)	0.023**
β2-mg (mg/L)	1.94(1.60-2.47)	2.29(1.83-2.80)	0.049**
Pre-CT	176.8 ± 40.5	160.5 ± 34.8	0.039**

CR: complete response; PR: partial response; SD: stable disease; CD: cumulative dose; DD: daily dose; Hb: hemoglobin; ALP: alkaline phosphatase; Alb: albumin; LDH: lactate dehydrogenase; Ca: calcium; Cr: creatinine; CCr: creatinine clearance rate; β2 mg: beta 2 microglobulin; Pre-CT: CT attenuation value before chemotherapy.

* p < 0.1.

** p < 0.05.

BMD following treatments

Most of the patients (17/18) who were diagnosed with osteoporosis before chemotherapy were still diagnosed with osteoporosis after chemotherapy and were not included. The remaining 79 patients were divided into two groups according to their CT attenuation values after chemotherapy. Osteoporosis occurred in 21 patients (26.6%). Table 4 shows the results of the univariate analysis. The two groups differed significantly based on age, BMI, and CD (p < 0.05). In the multivariate analysis for the occurrence of osteoporosis after chemotherapy, age (OR = 3.36, 95%CI 1.95 ∼ 5.81, p < 0.001) and CD (OR = 2.09, 95%CI 1.28 ∼ 3.40, p = 0.003) were found to be risk factors. There was no statistically significant correlation between osteoporosis and BMI (Figure 2c).

Table 4. Univariate analysis of BMD following treatment.

	Group Cpost(n = 58)	Group Opost (n = 21)	p-value
Age	45.1 ± 11.0	51.9 ± 8.6	0.013**
Sex			0.984
Male	25(43.1%)	9(42.9%)
Female	33(56.9%)	12(57.1%)
BMI	23.9(20.6 ∼ 25.7)	26.0(23.5 ∼ 27.8)	0.022**
Grade			0.531
I-II	43(74.1%)	17(81.0%)
III	15(25.9%)	4(19.0%)
Ann Arbor Stage			0.331
I-II	12(20.7%)	2(9.5%)
III-IV	46(79.3%)	19(90.5%)
B-symptom			0.414
N	44(75.9%)	14(66.7%)
Y	14(24.1%)	7(33.3%)
Bone marrow involvement			0.670
N	28(48.3%)	9(42.9%)
Y	30(51.7%)	12(57.1%)
FLIPI-2			0.856
0-1	39(67.2%)	16(76.2%)
2	10(17.2%)	3(14.3%)
≥ 3	9(15.5%)	2(9.5%)
Smoking history			0.527
N	48(82.8%)	16(76.2%)
Y	10(17.2%)	5(23.8%)
Calcium + Vitamin D			0.572
N	13(22.4%)	6(28.6%)
Y	45(77.6%)	15(71.4%)
Effect			0.205
CR	49(84.5%)	16(76.2%)
PR	7(12.1%)	2(9.5%)
SD	2(3.4%)	3(14.3%)
CD (mg)	2500(1500 ∼ 3025)	3000(2000 ∼ 4000)	0.029**
CD pre body surface (mg/m2)	1409(821 ∼ 1873)	1652(1165 ∼ 2396)	0.143
DD (mg/d)	15.9 ± 9.3	16.3 ± 6.2	0.835
Hb (g/L)	132(122 ∼ 153)	137(119 ∼ 149)	0.702
ALP (U/L)	69(59 ∼ 85)	65(55 ∼ 86)	0.633
Alb (g/L)	43.9(42.5 ∼ 46.0)	43.2(40.1 ∼ 46.8)	0.727
LDH (U/L)	197(174 ∼ 225)	191(169 ∼ 229)	0.824
Ca (mmol/L)	2.31 ± 0.13	2.28 ± 0.12	0.293
Cr (μmol/L)	73 ± 14	77 ± 17	0.351
CCr (ml/min)	95(83 ∼ 110)	97(78 ∼ 109)	0.618
β2-mg (mg/L)	2.18(1.59 ∼ 2.60)	2.15(1.79 ∼ 2.36)	0.934

Group Cpost: patients with normal BMD at the end of chemotherapy; Group Opost: patients with BMD ≤ 135HU at the end of chemotherapy; CR: complete response; PR: partial response; SD: stable disease; CD: cumulative dose; DD: daily dose; Hb: hemoglobin; ALP: alkaline phosphatase; Alb: albumin; LDH: lactate dehydrogenase; Ca: calcium; Cr: creatinine; CCr: creatinine clearance rate; β2 mg: beta 2 microglobulin;.

* p < 0.1.

** p < 0.05.

BMD trend following treatment

The follow-up data were divided into groups, and Table 5 shows the corresponding follow-up period and number of patients in each group. The differences in CT attenuation values at three points in each group were compared. The results of different follow-up times were consistent. The CT attenuation values before chemotherapy were higher than those after chemotherapy and those at follow-up, and there was no statistical difference between the latter two.

Table 5. Comparison at each follow-up point.

Point (month)	Number	Pre-CT	Post-CT	Follow–up
6 (6–9)	34	175.0 ± 41.3	154.7 ± 47.6	148.0 ± 48.8
12 (10–15)	16	171.1 ± 42.4	150.6 ± 42.1	143.0 ± 38.4
18 (16–21)	18	163.3 ± 30.5	141.0 ± 33.1	145.5 ± 36.4
24 (22–27)	17	174.4 ± 42.2	151.0 ± 46.9	152.9 ± 44.9
30 (28–33)	17	163.1 ± 30.2	138.8 ± 34.8	145.5 ± 40.0
36 (34–39)	8	167.9 ± 23.8	144.6 ± 28.8	146.0 ± 20.5
42 (≥ 40)	15	157.0 ± 21.9	131.0 ± 22.7	133.9 ± 20.9

The patient cohort with follow-up in a certain period was classified as one group, and the CT attenuation values of pre-chemotherapy, post-chemotherapy, and the follow-up point were compared.

CT images were available at 6 months follow-up and ≥ 12 months follow-up in 20 patients. CT attenuation values were examined at four different time points: pre-chemotherapy, post-chemotherapy, 6 months, and final follow-up, respectively. The median follow-up time was 25 (range, 13–69) months. The CT attenuation values at four time points were 171.4 ± 42.2HU, 154.2 ± 46.0HU, 145.0 ± 49.1HU, and 151.1 ± 53.7HU (Figure 3), with statistically significant variations (p < 0.001). The post hoc test findings revealed that pre-chemotherapy had the highest CT attenuation value, statistically different from the other three groups. Post-chemotherapy had a higher value than that of 6 months, and the difference was statistically significant (adjusted p < 0.05). The value of final follow-up was greater than that of 6 months but less than that of post-chemotherapy. Nonetheless, neither of them was statistically significant.

Figure 3. Trend of BMD CT attenuation values at four time points in 20 patients.

Pre: pre-chemotherapy; Post: post-chemotherapy; final: final follow-up. *: <0.05 **: <0.01 ***: <0.001

Discussion

Although most patients (71.4%) were treated with calcium and Vitamin D, they continued to experience BMD reduction and developed osteoporosis. Although some patients with high baseline BMD did not develop osteoporosis after treatment, we believe that identifying factors associated with decreased BMD can help us accurately identify high-risk patients and initiate anti-resorptive therapy early in treatment. To address the issue, we retrospectively analyzed factors impacting the occurrence of osteoporosis and leading to significant BMD reduction during chemotherapy, finally providing evidence for the prevention and treatment of osteoporosis, and decreasing mortality and disability post-treatment in patients with NHL.

The incidence of osteoporosis in patients with NHL managed by a wait-and-watch strategy was 10% [14], similar to that in the general population [15]. Another Danish cohort nationwide study also demonstrated that the incidence of osteoporotic events in untreated patients with FL was not remarkably increased compared to that in the normal population [16]. Our results showed that there was no relationship between the presence of osteoporosis at the time of FL diagnosis and the grade, stage, and FLIPI-2 score. There is currently no evidence to support that FL drives the development of osteoporosis, and the reasonable hypothesis is that the cause of osteoporosis is bone loss during chemotherapy.

To investigate the factors impacting bone loss in patients with FL, we retrospectively analyzed the differences in significant BMD reduction (>15%) between patients and controls. In a recent study by Paccou, 32 patients with lymphoma receiving chemotherapy experienced significant BMD reduction in the lumbar spine, where multivariate analysis showed that female sex, low CCr, and high lactate dehydrogenase (LDH) were risk factors [17]. Another study involving 213 patients with Hodgkin Lymphoma (HL) concluded that chemotherapy protocol and age were risk factors for significant reduction in BMD (>15%) [18]. Using the same cutoff as above, the proportion of patients with significantly decreased BMD in our analysis was 44.3%, and low baseline CT attenuation value and low Hb were risk factors.

Hb is used as one of the nutritional indicators that is frequently associated with chronic diseases. A few studies have assessed the association between low Hb and osteoporosis in other specific populations [19, 20], and anemia leading to osteoporosis has been mostly attributed to increased levels of erythropoietin and a compensatory expansion of bone marrow [21]. In contrast, bone health and bone marrow status interplay, and Hb is also one of the indicators for evaluating the prognosis of FL. Paccou et al. found that baseline Hb was associated with total hip BMD in univariate analysis but showed no statistical significance in multivariate analysis in patients with NHL [17]. The potential interacting mechanisms should be further explored.

According to Amiche’s study, excessive GC use (DD ≥ 15 mg and CD ≥ 1 g) significantly increases fracture risk [22]. The FRAX tool, a fracture risk assessment tool, revealed that the fracture risk rose as the DD increased [23]. However, we believe that using pulsed high-dose GCs in NHL treatment is different from the long-term use of low-dose GCs. The HL study showed a relationship between CD per body surface >3400mg/m2 and significant BMD reduction [18]. An analysis of the FLYER-trial showed BMD reduction was greater in patients with diffuse large B-cell lymphoma (DLBCL) receiving six cycles of RCHOP than in those receiving four cycles, with different CD (3000 mg versus 2200 mg) and the same DD values [11]. Therefore, we recommend CD as a better predictive index for osteoporosis following chemotherapy compared to DD.

We hypothesized that CD would be associated with both BMD reduction and osteoporosis occurrence. However, our study showed no association between CD and BMD reduction, which is similar to the findings of Paccou et al. [17]. Pedersen et al. showed that patients receiving RCHOP experienced a greater BMD loss during chemotherapy than those receiving RB, with a median CD of 3000 mg in the RCHOP group and 0 mg in the RB group [24]. We compared the two most commonly used protocols (RCHOP versus RCHOP + RFC); however, there was no significant difference in ΔBMD (Table S1). A possible reason is that limited dose differences could not lead to differences in BMD reduction.

BMD continued to decline after chemotherapy and did not return to the baseline. In the analysis of consecutive follow-up patients, BMD was lower than that at the end of chemotherapy during the 6 months of follow-up. Of note, data from only 20 patients satisfied the study of BMD trend of following treatment, which may have led to a bias. However, a similar trend was observed in patients treated with oral GCs, with a significant increase in fracture risk 3 months following treatment and a fast recovery to baseline 1 year after oral GC discontinuation [25]. Our results for long-term trends of BMD following treatment are similar to those of Pedersen et al. which showed that at 1 to 3 years follow-up, BMD was lower than baseline but not significantly different from the end of chemotherapy. They also showed a continued reduction in BMD at the 4- and 5-year follow-up; however, the authors did not discuss the reason, and we speculate that it may be a bias due to declined follow-up. We recommend BMD screening within 6 months following chemotherapy to confirm the lowest level of BMD, and patients at the two-year follow-up without osteoporosis can be screened as the general population [16].

We also found that the supplementation with calcium and Vitamin D did not significantly prevent BMD reduction and osteoporosis, and this may be because this supplementation was insufficient to counteract the effect of GCs on bone metabolism. Svendsen et al. also found that calcium and Vitamin D did not protect against BMD reduction [10], while the effect of bisphosphonate treatment on BMD decline in patients with NHL has been confirmed in prospective studies [26, 27]. Therefore, strategies for prevention and treatment of osteoporosis should be more focused on patients at greater risk.

This study had some limitations. First, because it was a retrospective analysis, there are bound to be some biases and inconsistencies in the patients included and their examinations. PET-CT results showed five patients (5%) with fractures, including four with rib fractures and one with a lumbar compression fracture, and this value was lower than the 14% documented in patients with DLBCL in a previous study [10]. Possibly, some patients did not undergo PET-CT following chemotherapy, resulting in an underestimation of fracture incidence. Second, CT attenuation values cannot replace DXA in diagnosing osteoporosis. CT attenuation values are currently only used as an opportunistic screening tool rather than a tool to guide treatment, and the osteoporosis threshold varies between studies. Third, the number of patients was relatively limited, and this prevented subgroup analysis and the inclusion of more factors in the multivariate analysis. We plan to include more patients with NHL in future studies.

Conclusions

Following chemotherapy, patients with FL experienced BMD reduction. The baseline CT attenuation value and baseline Hb levels were risk factors for significant BMD reduction. Age and CD impacted the occurrence of osteoporosis after chemotherapy. The effect of chemotherapy on BMD reduction persisted for 6 months and BMD values never returned to baseline levels. Therefore, BMD should be monitored for 6 months post chemotherapy. BMD evaluation and more positive anti-resorptive treatment should be administered for such high-risk patients.

Ethical approval

Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Medical Science Research Ethics Committee (No.M2021547).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This study was funded by Peking University Third Hospital Cohort Study (BYSYDL2021011)