Osteoporosis International

, Volume 17, Issue 2, pp 273–280

Reduced risk of back pain following teriparatide treatment: a meta-analysis

Authors

  • Michael C. Nevitt
    • University of California San Francisco
  • Peiqi Chen
    • Lilly Research LaboratoriesEli Lilly and Company
  • Robin K. Dore
    • University of California Los Angeles
  • Jean-Yves Reginster
    • University of Liège
  • Douglas P. Kiel
    • Hebrew Rehabilitation Center for the Aged
    • Harvard Medical SchoolHarvard University
  • Jose R. Zanchetta
    • Instituto de Investigaciones Metabólicas
  • Emmett V. Glass
    • Lilly Research LaboratoriesEli Lilly and Company
    • Lilly Research LaboratoriesEli Lilly and Company
Original Article

DOI: 10.1007/s00198-005-2013-2

Cite this article as:
Nevitt, M.C., Chen, P., Dore, R.K. et al. Osteoporos Int (2006) 17: 273. doi:10.1007/s00198-005-2013-2

Abstract

Vertebral fractures are the most common osteoporotic fracture and may result in back pain with functional limitations and diminished quality of life. Teriparatide [rhPTH (1–34)] has been shown to increase bone mass and reduce the risk of vertebral and other osteoporotic fractures. The aim of this study was to evaluate the effects of teriparatide on the risk of back pain in patients with osteoporosis. A systematic review of the literature was performed, and five trials were identified and included in our analyses. All trials were randomized, double-blinded, and parallel with either new vertebral fracture ( n =1) or bone mineral density as the primary endpoint ( n =4). Four studies were in postmenopausal women with osteoporosis, and one was in men with idiopathic or hypogonadal osteoporosis. Two trials were placebo controlled, two trials were alendronate controlled, and one trial involved teriparatide plus hormone replacement therapy versus hormone replacement therapy alone. Reports of back pain, defined as new or worsened back pain after initiating the study drug, were obtained from adverse event databases, and the risk of back pain was analyzed using a multivariate Cox proportional hazards model. Results were not statistically heterogeneous ( P =0.60) across trials, and there were no differences between groups administered teriparatide 20 or 40 mcg/day doses ( P =0.64). The rates of back pain, moderate or severe back pain, and severe back pain per 100 patient-years were numerically lower in the teriparatide versus comparator groups in each study. Compared with the pooled comparator, patients in the pooled teriparatide group had reduced risk for any back pain [relative risk, 0.66 (95% CI, 0.55–0.80)], moderate or severe back pain [relative risk, 0.60 (95% CI, 0.48–0.75)] and severe back pain [relative risk, 0.44 (95% CI, 0.28–0.68)]. Separate meta-analyses comparing teriparatide versus placebo or antiresorptive drugs gave similar results. In conclusion, patients randomized to teriparatide had a reduced risk of new or worsening back pain compared to patients randomized to placebo, hormone replacement therapy or alendronate.

Keywords

AlendronateBack painHormone replacement therapyOsteoporosisrhPTH 1–34Teriparatide

Introduction

Vertebral fractures are the most frequent outcome in osteoporosis with an annual occurrence of 700,000 in the United States [1]. Approximately 30% of women will have sustained a vertebral fracture by the age of 75, and 50% will have sustained a vertebral fracture by the age of 85 [2]. The incidence of vertebral fractures is expected to increase as the population aged 85 and older is expected to double by 2020 [3]. The majority (84%) of clinically diagnosed vertebral fractures results from the investigation of back pain [4]. Back pain resulting from vertebral fracture often has an overwhelming effect on women with osteoporosis [5, 6, 7]. The risk of chronic severe back pain increases with the number and severity of vertebral fracture [8, 9]. In addition to severe chronic back pain, vertebral fractures result in patient depression, disability, loss of productivity and independence, and reduced quality of life [2, 10, 11, 12, 13]. Back pain itself appears to contribute more to disability than does vertebral fracture [14].

The direct cost of treating osteoporotic fractures in the United States was estimated to be $13.8 billion in 1995 [15]. Average hospital costs for treating vertebral fractures are 50–63% that of hip fractures and account for 25% of annual hospital admissions related to osteoporosis. The majority of these costs are associated with the treatment of back pain and back-related disability [16, 17]. In clinical practice, acute back pain is managed by analgesics and nonpharmacological modalities, such as physical strengthening [13, 18], and in cases of severe chronic back pain, surgical vertebral augmentation techniques are considered [19, 20]. Publications from large, randomized, controlled trials of antiresorptive agents for the treatment of osteoporosis have generally not included any mention of back pain [21, 22, 23, 24, 25, 26].

Teriparatide [rhPTH (1–34)] was shown in the double-blind randomized Fracture Prevention Trial to reduce the risk of vertebral fractures in women with postmenopausal osteoporosis [27]. Additionally, back pain was reported to occur in 23% of women treated with placebo versus 17% of women treated with teriparatide 20 mcg and 16% of women treated with teriparatide 40 mcg. In a separate randomized, double-blind study of women with osteoporosis, 5.5% of women randomized to teriparatide 40 mcg reported back pain compared with 19.2% of women randomized to alendronate 10 mg [28]. We conducted a meta-analysis of randomized, double-blind, controlled trials of teriparatide to test the hypothesis that teriparatide reduces the risk of developing back pain when compared to placebo or other comparator study drugs.

Materials and methods

Inclusion criteria for meta-analysis

We performed a systematic review of the literature [Pubmed and Medline (1966 to March 2005), Embase 1988 to 2005 week 13] combining medical subject headings with search terms (teriparatide, PTH 1–34, randomized, controlled trials and osteoporosis) and identified three publications that met our inclusion criteria of being randomized, blinded, clinical trials of PTH (1–34) versus a comparator group with collection of back pain data. Nine studies of daily administration of other formulations and preparations of PTH were also identified, but none of these studies included data on back pain in published reports, and one study of weekly PTH 1–34 lacked a non-PTH control group [29]. Therefore, only trials of teriparatide [rhPTH (1–34)] were included in analyses since back pain data were not available in studies of other PTH (1–34) preparations. Two additional studies were identified from data on file at Eli Lilly and Company, and both of these met our inclusion criteria. The five studies included in these analyses included: two placebo-controlled trials [27, 30], two alendronate 10 mg daily comparator trials [28, 31], and one study of hormone replacement therapy (HRT) versus HRT plus teriparatide [32]. All trials were randomized, double-blinded, and parallel with either new vertebral osteoporotic fracture ( n =1) or bone mineral density as the primary endpoint ( n =4).

Study participants

Patients were eligible for enrollment in these clinical trials if they were free of chronic disabling diseases other than osteoporosis. Women were required to be at least 5 years postmenopausal. Previous vertebral fracture status and lumbar spine or total hip bone mineral density (BMD) inclusion requirements are given in Table 1. Men were diagnosed with idiopathic or hypogonadal osteoporosis. Patients were randomized to receive teriparatide 20 mcg or teriparatide 40 mcg by once-daily subcutaneous self-injection or comparator therapy. All patients received daily calcium (1,000 mg) and vitamin D (400–1,200 IU) supplements. Patients with any disease known to affect bone or calcium metabolism, urolithiasis within 2 years, serum creatinine levels greater than 2 mg/dl, alcohol or drug abuse, or medication use known to alter bone metabolism within the previous 2 to 24 months depending on the drug were excluded from participation in the trials. In all trials, Institutional Review Board approval was obtained at all study sites and written informed consent was obtained from each patient. All trial methods and procedures were conducted according to the ethical standards of the Declaration of Helsinki.
Table 1

Overview of studies and baseline demographics of trial participants. All trials were multicenter, double-blind, parallel randomized-controlled trials. Patients in all trials were given calcium and vitamin D supplements. TPTD20 teriparatide 20 mcg/day; TPTD40 teriparatide 40 mcg/day; ALN10 alendronate 10 mg/day; HRT hormone replacement therapy

Comparison (reference)*

Gender

Inclusion criteria

Median observation (months)

Group (n)

Mean age (years ± SD)

LS BMD (g/cm2±SD)

A. Phase 3, placebo versus TPTD20 versus TPTD40 [27]

Aged 30–85 years, at least 5 years postmenopausal, at least one moderate or two mild atraumatic VFX. Hip or spine BMD T-score ≤–1.0 if fewer than two moderate fractures

20.1

PL (544)

69±7

0.82±0.17

TPTD20 (541)

70±7

0.82±0.17

TPTD40 (552)

70±7

0.82±0.17

B. Phase 3, placebo versus TPTD20 versus TPTD40 [30]

Aged 30–85, idiopathic or hypogonadal osteoporosis, hip or spine T-score ≤–2.0

11.5

PL (147)

59±13

0.85±0.14

TPTD20 (151)

59±13

0.89±0.15

TPTD40 (139)

58±13

0.87±0.14

C. Phase 3, ALN10 plus placebo injection versus placebo pill plus TPTD40 injection [28]

Aged 35–85, at least 5 years postmenopausal, femoral neck or spine T-score ≤−2.5

15.4

ALN10 (73)

65±9

0.80±0.12

TPTD40 (73)

66±8

0.80±0.11

D. Phase 3, HRT plus placebo injection versus HRT plus TPTD40 injection [32]

Aged 30–85, at least 5 years postmenopausal, hip or spine T-score ≤–1.0

15.0

HRT (125)

61±7

0.91±0.15

TPTD40 + HRT (122)

62±8

0.90±0.15

E. Phase 4, ALN10 plus placebo injection versus placebo pill plus TPTD20 injection [31]

Aged 45–85 years, at least 5 years postmenopausal, hip or spine T-score between −2.5 and −4.0 inclusive

17.6

ALN10 (101)

67±9

0.75±0.10

TPTD20 (102)

65±8

0.75±0.08

Assessment of back pain

Back pain data were collected at each visit in each trial during the monitoring of adverse events. Investigators did not initiate queries on back pain. Back pain was defined as either new or worsened in severity after the initiation of the study drug. Patients who did not spontaneously report back pain as a preexisting health condition prior to treatment initiation were eligible to develop new back pain by adverse event reports. Worsened back pain was defined as the first increase in back pain severity after initiating the study drug in patients with preexisting back pain at baseline. Therefore, patients with baseline mild back pain were only at risk for moderate or severe back pain, and patients with baseline moderate back pain were only at risk for severe back pain. The moderate or severe back pain classification included patients with either moderate or severe back pain, while the severe back pain classification included only patients with severe back pain. For each patient, investigators were instructed to list all preexisting conditions or symptoms present at entry to the study, all clinically relevant abnormalities found on the physical exam, electrocardiograms or X-rays, and all adverse events that occurred during the study on standardized “running records” case report forms. Information collected included a description of the condition/event, the COSTART class term, onset and stop date, an indication if the adverse event was serious, the severity by visit and relationship to the study drug. Study investigators were given instructions for the assignment of adverse event severity as mild, moderate or severe. Mild was defined as no change in physical activity with occasional medication use for relief of symptoms. Moderate included mild disruptions in daily physical activities and regular medication use for the alleviation of symptoms. Criteria for severe were major disruptions in normal daily activities, additional medication use and additional treatment above and beyond normal that may have included hospitalization. Standard directions were provided to all investigators to facilitate the uniform collection of adverse events in these trials.

Statistical analyses

Data were analyzed using an intention-to-treat approach using individual level data made available by the research laboratories of Eli Lilly and Company. The number of patient-years at risk for back pain was calculated for each patient as defined by the number of years measured from the date of randomization to the first occurrence of back pain or the last date of participation in the study. A multivariate Cox proportional hazard model with the study as a stratification factor was used to compute the relative risk of back pain (pooled teriparatide versus pooled comparator). Relative risks of the individual studies were also computed using the Cox proportional hazard model without study as a stratification factor. The interaction of treatment (teriparatide versus placebo or other study drug) and study (using indicator variables for the studies) in a Cox model was used to assess the heterogeneity of results across trials with an alpha level of less than 0.1 considered statistically significant. To ensure that the results of the meta-analysis were not influenced by any particular trial, a sensitivity analysis was performed by serially repeating the analyses after exclusion of each individual trial. The cumulative proportion of patients with any new or worsening back pain was calculated using the Kaplan-Meier method for pooled teriparatide and comparator groups. All statistical tests were performed using SAS software, version 8.2 (SAS Institute, Inc., Cary, N.C.).

Results

A summary of the trials and baseline demographics of the trial participants is shown in Table 1. The review identified five randomized, controlled trials with a total of 2,670 patients. The methods of randomization and blinding and details regarding dropouts and withdrawals were described completely in all five trials. The trials included dissimilar study enrollment criteria for lumbar spine and hip T-scores and baseline fracture status.

The results of the meta-analyses are shown in Table 2 and Fig. 1A–C. The rate of any back pain, moderate or severe back pain, and severe back pain per 100 patient-years at risk was 12.05, 7.63 and 1.54 in the pooled teriparatide group compared to 18.59, 12.78 and 3.62 in the pooled comparator group, respectively (Table 2). Compared to the pooled comparator group, patients in the pooled teriparatide group had a reduced risk for developing any back pain [relative risk, RR, 0.66 (95% CI, 0.55 to 0.80)], moderate or severe back pain [RR, 0.60 (95% CI, 0.48 to 0.75)] and severe back pain [RR, 0.44 (95% CI, 0.28 to 0.68)] (Fig. 1A–C). In a meta-analysis of pooled placebo comparator trials (studies A and B), teriparatide-treated patients had a significantly reduced risk of any back pain [RR, 0.76 (95% CI, 0.57 to 0.88)], moderate or severe back pain [RR, 0.67 (95% CI, 0.52 to 0.87)] and severe back pain [RR, 0.47 (95% CI, 0.28 to 0.78)] (Fig. 1A–C). Similarly, a meta-analysis of pooled antiresorptive comparator trials (studies C, D and E) showed significant reductions in teriparatide-treated patients for any back pain [RR, 0.54 (95% CI, 0.36 to 0.81)], moderate or severe back pain [RR, 0.40 (95% CI, 0.24 to 0.67)] and severe back pain [RR, 0.33 (95% CI, 0.12 to 0.90)], respectively (Fig. 1A–C).
Table 2

The number of cases, patient-years at risk, and rate of any back pain, moderate or severe back pain and severe back pain by study. TPTD20 teriparatide 20 mcg/day; TPTD40 teriparatide 40 mcg/day; ALN10 alendronate 10 mg/day; HRT hormone replacement therapy

Study

Treatment group

No.

Any back pain

Moderate or severe back pain

Severe back pain

Cases

Patient-years at risk

Rate (per 100 patient-years)

Cases

Patient-years at risk

Rate (per 100 patient-years)

Cases

Patient-years at risk

Rate (per 100 patient-years)

A

Placebo

544

123

700.7

17.6

90

749.8

12.0

28

823.1

3.40

A

TPTD20

541

91

725.3

12.6

62

765.3

8.10

12

819.1

1.47

A

TPTD40

552

87

727.2

12.0

61

764.5

7.98

16

815.0

1.96

B

Placebo

147

19

128.0

14.8

12

133.1

9.02

4

138.6

2.89

B

TPTD20

151

14

127.6

11.0

10

129.8

7.70

0

136.0

0.00

B

TPTD40

139

11

113.2

9.71

5

117.1

4.27

1

119.6

0.84

C

ALN10

73

14

72.0

19.4

10

75.2

13.3

2

81.3

2.46

C

TPTD40

73

4

74.6

5.36

2

76.6

2.61

1

76.6

1.31

D

HRT

125

13

131.0

9.85

7

137.3

5.10

1

142.4

0.70

D

TPTD40 HRT

122

6

126.5

4.74

4

128.4

3.12

0

132.0

0.00

E

ALN10

101

39

86.0

45.4

33

93.8

35.2

12

113.8

10.5

E

TPTD20

102

26

89.0

29.2

15

102.7

14.6

4

113. 0

3.54

Pooled

Comparator

990

208

1,118.7

18.6

152

1,189.1

12.8

47

1,299.3

3.62

Pooled

TPTD

1,680

239

1,983.5

12.1

159

2,084.5

7.6

34

2,211.3

1.54

Fig. 1

Relative risk of any back pain ( A), moderate or severe back pain ( B) and severe back pain ( C) stratified by treatment group and study

Kaplan-Meier analysis of the cumulative percent of patients experiencing any back pain in pooled teriparatide and comparator groups (Fig. 2) showed that the curves appeared to separate after approximately 3 months with overall statistically significant ( P =0.001) divergence of back pain between groups over the entire time course. After 21 months, the cumulative percent of patients experiencing any back pain was approximately 16% in the pooled teriparatide group compared to 23.5% in the pooled comparator group.
Fig. 2

Cumulative percent of patients with any back pain. * P -value from log-rank test for the divergence of back pain between groups over entire time course

Back pain results were not statistically heterogeneous ( P =0.60) across trials, and there were no differences between groups administered teriparatide 20 or 40 mcg/day doses ( P =0.64). Lack of heterogeneity allowed for teriparatide groups to be pooled and compared versus pooled comparator groups. Sensitivity analysis results demonstrated the reductions in the risk of back pain were consistent across trials and were robust to the removal of each individual trial from the meta-analysis (Table 3).
Table 3

Estimates for the relative risks for any back pain, moderate or severe back pain, and severe back pain when individual trial was deleted from the meta-analysis

Deleted study

Any back pain

Moderate or severe back pain

Severe back pain

A

0.58

0.47

0.27

(0.42–0.81)

(0.31–0.71)

(0.11–0.68)

B

0.66

0.60

0.47

(0.54–0.80)

(0.47–0.75)

(0.30–0.74)

C

0.68

0.62

0.44

(0.56–0.83)

(0.50–0.78)

(0.28–0.69)

D

0.67

0.60

0.44

(0.55–0.81)

(0.48–0.75)

(0.28–0.70)

E

0.66

0.64

0.46

(0.54–0.81)

(0.50–0.82)

(0.28–0.76)

None

0.66

0.60

0.44

(0.55–0.80)

(0.48–0.75)

(0.28–0.68)

Discussion

This meta-analysis demonstrates that compared to placebo, alendronate, or hormone replacement, teriparatide-treated women had a significantly reduced risk of back pain, moderate or severe back pain, and severe back pain. The reduction in the risk of back pain endpoints was not heterogeneous across trials. Furthermore, there was no significant difference between teriparatide 20 mcg per day and teriparatide 40 mcg per day; these findings justified pooling the comparator groups and pooling the teriparatide dose groups. Compared to the pooled comparator group, patients in the pooled teriparatide group had a significantly reduced risk for any back pain, which became evident after 3 months of the study drug. Additionally, moderate or severe back pain and severe back pain were significantly reduced in the pooled teriparatide versus the pooled comparator group. The significant reductions in all back pain endpoints in our sensitivity analyses indicated that the overall findings were not driven by any single trial. Additional meta-analyses of the two teriparatide versus placebo trials and the three teriparatide versus antiresorptive trials also showed a reduced risk of back pain, moderate or severe back pain, and severe back pain in the teriparatide-treated patients.

Importantly, the back pain data included in these analyses were from the routine adverse event monitoring in these trials. Patients were not queried specifically regarding back pain in these trials. The relevance of this adverse event data collection is that it may be somewhat similar to that employed in clinical practice, and if the results are correct, the implication is that patients treated with teriparatide should less often spontaneously report back pain of various severities compared to otherwise similar patients treated with placebo, alendronate or hormone replacement therapy. The approach used to assess back pain in these studies is particularly well-suited for episodes of acute but transient back pain that typically result from new vertebral fractures.

Back pain has numerous possible causes. In women with osteoporosis, vertebral fracture is an especially important etiology, and the incidence of back pain increases with increasing incident vertebral fracture severity [33]. The overall results show that women treated with teriparatide had a marked reduction in overall back pain and more severe back pain that would be associated with decreased activities [14].

Only one trial (A) included both baseline and follow-up radiographs [27]. Compared with placebo, women in the teriparatide group had a reduced risk for back pain with a subsequent finding of new vertebral fractures, more severe vertebral fractures and multiple vertebral fractures [34]. In the teriparatide 20 mcg group, fewer patients reported new or worsening back pain during follow-up and were also found to have one or more new vertebral fractures during follow-up (5 of 444, 1%) than in the placebo group (29 of 448, 6%) [34]; this 5% absolute risk difference is similar to the 6% absolute risk difference in the incidence of any new or worsening back pain between the groups [27]. The reduced risk of back pain in teriparatide-treated patients is thus consistent with a reduction in the risk of new painful vertebral fractures. Additionally, teriparatide has been shown in animal models to accelerate fracture healing [35, 36], and it is possible that teriparatide heals and stabilizes pre-existing fractures or microdamage in patients with advanced osteoporosis. There are no data suggesting an analgesic effect of teriparatide. The inclusion of radiographs in only one trial (A) precludes analyses of the mechanism of back pain reduction in the other studies.

The reduced risk of back pain in women treated with teriparatide versus alendronate and hormone replacement therapy is noteworthy because these drugs are all known to prevent vertebral fractures. However, while teriparatide has been shown to reduce the overall risk of back pain, large trials of antiresorptive drugs, including trials of alendronate, risedronate, raloxifene and calcitonin, have generally not included any mention of back pain [21, 22, 23, 24, 25, 26]. Nevitt et al. published back pain results from the Fracture Intervention Trial study of women with prevalent vertebral fractures [37]. Compared to placebo, significantly fewer alendronate-treated patients required bed-rest because of back pain and limited their activity because of back pain. Those taking alendronate had substantial reductions in the number of days of bed-rest for back pain and days of limited activity due to back pain. The authors attributed these effects on back pain to the reduction in new vertebral fractures among women treated with alendronate. However, statistically significant differences were not observed between treatment groups in the number of patients with any back pain or limited activity due to back pain. Thus, back pain outcomes in teriparatide compared with alendronate-treated women might be due to a greater effect relative to alendronate on vertebral fracture-related back pain prevention. It is noteworthy that hormone replacement therapy has been reported to increase back pain. The mechanism by which hormone replacement therapy might increase back pain is postulated to be an increase in joint laxity or secondary effects on connective tissue leading to a higher occurrence of back pain in certain groups of women [38, 39].

In conclusion, back pain risk reduction in patients treated with teriparatide therapy was consistent across trials. Patients randomized to teriparatide had a reduced risk of developing new or worsening back pain compared to patients randomized to comparators. The relative risk of back pain was reduced in teriparatide-treated patients compared with both placebo and antiresorptive-treated patients. The consistency of the findings across trials suggests that teriparatide reduces back pain risk and that bone-building therapy with teriparatide may have different effects on back pain risk versus antiresorptive therapies. Confirmation of these findings will require a prospective trial with back pain as the endpoint.

Acknowledgments

We are grateful to Mary Ellen Perron and Mindy Rance for assistance in the preparation of the figures. Funding was provided by Lilly Research Laboratories, Eli Lilly and Company.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2005