Osteoporosis International

, Volume 17, Issue 3, pp 478–483

Osteoporosis intervention in ambulatory patients with previous hip fracture: a multicentric, nationwide Italian survey

Authors

  • Vincenzo Carnevale
    • Department of Internal Medicine“Casa Sollievo della Sofferenza” Hospital IRCCS
  • Luciano Nieddu
    • “San Pio V” University
  • Elisabetta Romagnoli
    • Department of Clinical Sciences“La Sapienza” University
  • Elisabetta Bona
    • Department of Statistics“La Sapienza” University
  • Sara Piemonte
    • Department of Clinical Sciences“La Sapienza” University
  • Alfredo Scillitani
    • Department of Endocrinology“Casa Sollievo della Sofferenza” Hospital IRCCS
    • Department of Clinical Sciences“La Sapienza” University
Original Article

DOI: 10.1007/s00198-005-0010-0

Cite this article as:
Carnevale, V., Nieddu, L., Romagnoli, E. et al. Osteoporos Int (2006) 17: 478. doi:10.1007/s00198-005-0010-0

Abstract

Our study investigated the patterns of treatment and adherence to prescribed therapies in 2,191 ambulatory patients with previous hip osteoporotic fractures at 207 participating orthopedic centers throughout Italy. All patients who came to the attention of the involved orthopedic surgeons were administered a questionnaire investigating: age, sex, height, weight, date of admission and length of stay in the hospital, other previous clinical fractures, bone density or biochemical testing concerning mineral metabolism, treatment with bone-active drugs in the six months before the fracture, treatment after discharge from the hospital, continuous use of prescribed drugs, pain at the site of hip surgery, and comorbidity. A multivariate logistic regression model was applied, considering a subset of the variables in the questionnaire, in order to determine the factors that significantly influenced discontinuation of treatment after hip fracture. Among the patients, 88.1% were female and 86.2% of the subjects were older than 65. The mean length of hospital stay for hip fracture was 19.0±25.3 days. At the time of interview, the mean time elapsed since hospitalization was 542.9±1,197.3 days. A previous clinical fracture was referred by 20.2% of patients. Before hip fracture occurrence, 52.8% of patients had never received any kind of treatment, and this figure reached 80% if we also included those who had taken only calcium and/or vitamin D. Corresponding proportions after fracture were 22% and 31.3%, respectively. Finally, 52% of patients had stopped treatment given for osteoporosis after a mean period of 1.4 years. According to the results of the logistic regression, increasing age, pain [odds ratio (OR): 1.36; 95% confidence interval (CI): 1.21–1.65] and no use of diagnostic tests (OR: 2.46; CI: 1.79–3.37) showed a positive effect on the probability of quitting the medication. On the other hand, being female reduces by half (OR: 0.49; CI: 0.37–0.45) the probability of quitting medication. Our data showed a low rate of primary prevention, a still insufficient post-fracture therapy, along with a high rate of early discontinuation of osteoporosis medication in patients with previous hip fracture.

Keywords

Hip fractureOsteoporosis therapySecondary preventionSurvey

Introduction

Osteoporosis has been appropriately defined a silent epidemic, since about 50% of women who are 50 years of age or older will suffer an osteoporotic fracture in their lifetime [1]. Moreover, about one third of the hip fractures which occur in elderly U.S. residents occur in men [2]. This already ominous picture will inevitably worsen in the following years because of the increasing proportion of older individuals in the general population [3]. Due to the magnitude of the problem, different guidelines may give differing recommendations for fracture-prevention therapy, but they all agree that older individuals with previous osteoporotic fractures should receive treatment.

Among osteoporotic fractures, those of the hip are associated with the highest morbidity and mortality, with about 20% of patients dying and over 50% becoming institutionalized within the first year after fracture [4]. Therefore, hip fracture exerts a remarkable impact on the health care system [5] and is responsible for staggering human and economic costs. Furthermore, in patients with prior hip fractures, the risk of future vertebral fractures is increased by 2.5 times, and that of subsequent hip fracture is at least doubled [6, 7]. In this scenario, even a small decrease in relative fracture risk may bring about a significant reduction in the human and economic burden of osteoporosis on society.

Despite the availability of effective treatments for osteoporosis [8], several studies showed a significant gap in implementing therapeutic guidelines, even in patients with previous fractures [915]. On the other hand, most of these studies either investigated relatively small samples of patients, involved only one or a few centers, potentially included fractures from other causes besides osteoporosis, or did not assess whether the patients fulfilled the prescription or discontinued the drugs.

Our study was aimed at evaluating the adequacy of treatment and adherence to prescribed therapies in patients with previous osteoporotic hip fractures examined in 207 participating orthopedic centers throughout Italy.

Materials and methods

The investigation was carried out during the period of May–July 2003. It was conducted at 207 orthopedic surgeries all over Italy, following the approval of the local ethical committee. During this period, the doctors were asked to administer a questionnaire to all ambulatory patients who came to their attention with a previous hip osteoporotic fracture. The questionnaire was targeted at obtaining information about: age, sex, height, weight, date of admission to the hospital for the previous hip fracture, length of stay in the hospital, other previous clinical fractures, previous mass measurements (either by DEXA or ultrasound), previous biochemical tests concerning mineral metabolism (that is, calcium, phosphorus, total alkaline phosphatase, and creatinine), treatment with bone-active drugs in the six months before the fracture, treatment after discharge from the hospital, continuous use of the prescribed drugs, pain at the site of hip surgery, and comorbidity. Less than 10% of patients refused to fill in the questionnaire. A total of 2,219 patients were investigated. Questionnaires filled in by 28 of them could not be included in the final analysis, mainly because of the lack of essential data. Therefore, a total sample of 2,191 patients remained. We calculated the proportion of people who took at least one prescription for an osteoporosis medication in the six months before and after the fracture occurred. Anti-osteoporotic drugs included in this study were those marketed in Italy at that time, i.e., calcium, clodronate, estrogens, raloxifene, alendronate 10 mg, alendronate 70 mg, risedronate 5 mg, and vitamin D and its derivatives. For the sake of simplicity, all the patients taking calcium, vitamin D derivatives, or both drugs were considered as one group. The proportion of patients who continued to take the drugs at the time of investigation was also calculated. Finally, a multivariate logistic regression model was applied to the 2,191 patients, considering a subset of the variables in the questionnaire, in order to determine the factors that significantly influence the discontinuation of treatment after hip fracture. In this particular model, age was not considered as a continuous variable, since it was categorized into three age groups (≤73 years, 74–81 years, 82> years). We also studied the behavior of those who discontinued therapy with respect to age and sex. In order to do so, we considered the number of patients who discontinued the therapy at various time intervals (expressed in classes of different lengths). To take into account the effect of different length time intervals on the number of patients who discontinued treatment, we considered not the frequencies but, instead, their densities, obtained by dividing the original frequencies by the length of the class. Comparisons between sexes and among age classes have been performed using the χ2 test.

Results

We analyzed the data of 2,191 patients with hip fracture. Table 1 shows the main demographic and anthropometric data of the 2,191 investigated patients. As expected, there was a preponderance of female patients who reported a hip fracture (88.1%), so that the ratio between the two sexes was 7.4. Considering the sample as a whole, 86.2% of the investigated subjects were older than 65 and 57.0% of them were older than 75.
Table 1

Demographic and anthropometric data of patients studied (mean±SD)

 

Whole

Females

Males

N

2,191

1,931

260

Age (years)

76.1±10.0

77.1±11.4

76.0±9.8

Height (cm)

161.0±9.5

159.8±9.1

169.2±7.8

Weight (kg)

65.3±11.5

64.3±11.3

72.3±10.0

Body mass index (kg/m2)

25.1±3.9

25.1±4.0

25.1±3.1

At the time of interview, the average period elapsed since the hospitalization was 542.9±1197.3 days. The mean length of stay in the hospital for the fracture that the patients had was 19.0±25.3 days. Furthermore, 20.2% of patients referred had had a previous clinical fracture (data not shown).

The use of osteoporosis medication by the patients investigated during the 6 months before and after the fracture occurred is reported in Fig. 1. It is interesting to see that 52.8% of patients had never taken any kind of treatment before fracture occurrence, and that this figure reaches 80% if we also include those who had taken only calcium and/or vitamin D. Corresponding proportions after fracture are 22% and 31.3%, respectively. Finally, at the time of the interview (that is, after a mean period of 1.4 years), only 820 patients out of 1,703 (48.1%) were continuing some kind of treatment for osteoporosis. Figure 2 illustrates the distribution of patients by the use of drugs for osteoporosis. Patients have been stratified in relation to the use of drugs before and after fracture. It is important to note that 21.7% of patients stopped treatment after fracture (white box, left); furthermore, 60.6% of patients continued not to take any drugs after fracture (white box, right).
https://static-content.springer.com/image/art%3A10.1007%2Fs00198-005-0010-0/MediaObjects/198_2005_10_Fig1_HTML.gif
Fig. 1

Use of osteoporosis medication after hip fracture, stratified in relation to the use of drugs before fracture. HRT indicates estrogen replacement therapy

https://static-content.springer.com/image/art%3A10.1007%2Fs00198-005-0010-0/MediaObjects/198_2005_10_Fig2_HTML.gif
Fig. 2

Use of osteoporosis medication during the six months before and after hip fracture

As far as the percentage of patients who stopped medications after fracture had occurred, these were 71.7% for those who were taking raloxifene, 64.6% for alendronate daily, 43.6% for risedronate daily, and 13.2% for alendronate weekly.

The results obtained from the application of a logistic regression model to the dataset available are reported in Table 2. The aim of this analysis was to determine the factors that, in our cohort, significantly influence (95% significance level) the continuation (0) or the discontinuation (1) of the use of medications after fracture. The influence of the covariates age, body mass index, the number of other pathologies, and the factors sex (male/female), diagnostic tests (yes/no), biochemical tests (yes/no), pain (yes/no), and other fractures (yes/no) on the probability of quitting the use of drugs has then been studied. Only sex, pain, age, and diagnostic tests show a significant effect on the probability of quitting the cure. More precisely, diagnostic tests, pain, and age show a positive effect on the probability of quitting medication. A patient who has not undergone diagnostic tests is more than twice as likely (2.46) to stop medication; the probability of quitting also increases by 1.36 if the patient experiences pain. Regarding age, if we consider the oldest age class as the constant one, people under 73 years have half the probability (0.57) of quitting the cure and the probability increases for the 74–81 year group. In other words, the probability of discontinuing treatment increases as people get older. On the other hand, being female reduces by half (0.49) the probability of quitting medication.
Table 2

Factors associated with the use of medication for osteoporosis after a hip fracture

Predictor variable

Odds ratio (OR)

95% confidence interval (CI)

Diagnostic test (no)

2.46

1.79–3.37

Pain (yes)

1.36

1.12–1.65

Age ≤73 years

0.57

0.45–0.72

74–81 years

0.69

0.54–0.87

Female sex

0.49

0.37–0.65

Figure 3 shows dropout curves with time of those patients who stopped therapy, with respect to sex (upper part) and age in classes (lower part). There were no statistical differences between the sexes concerning the behavior of those who quit therapy (χ2=9.450; p=0.09). However, considering the pattern with respect to age, there was a significant statistical difference. Indeed, the suspension of therapy, mostly noticed in the immediate period following fracture, was most common in the oldest with respect to youngest patients (χ2=39.985; p=0.00). This was also confirmed by the correlation coefficient among the age and the time period in which the therapy was discontinued, which is negative and statistically different from zero (χ2=–0.145; p=0.00).
https://static-content.springer.com/image/art%3A10.1007%2Fs00198-005-0010-0/MediaObjects/198_2005_10_Fig3_HTML.gif
Fig. 3

Dropout curves with time of those patients who stopped therapy, with respect to sex (upper part) and age in classes (lower part)

Discussion

Our data provide relevant information about the current management of patients who have had a fragility hip fracture in our country. In fact, they are virtually all admitted to hospital care, which, in turn, gives them the unique opportunity to receive diagnosis and treatment for osteoporosis. The latter is of the utmost importance for patients who survive their hip fractures, whose risk of future hip fracture is two to six times higher than the general population [6, 7].

Our results were obtained from a very large sample of patients of both sexes who experienced the same devastating osteoporotic fracture. The cross-sectional multicentric design of the study allowed us to collect, in a short time frame, data from all over Italy. This way, our results were not influenced by potential pitfalls resulting from either different local attitudes or varying historical perspectives to the disease. The high response rate we obtained should quite likely be ascribed to the fact that the questionnaires were directly administered by the attending physicians. This approach also allowed us to identify fragility fracture by clinical judgment, and, thus, avoided some misclassification bias invariably present in series which rely only on ICD codes for diagnoses. The latter, in fact, besides possible coding inaccuracies, could also include high-force fractures (e.g., closed fractures from motor vehicle crash), which are not real osteoporotic ones [10, 13, 14]. The mean time elapsed from hip fracture events in our patients was long enough to verify not only the prescriptions they received both at discharge and from primary care physicians, but also their long-term compliance.

In our study, one in every five patients referred had a former clinical fracture. Nevertheless, despite such an ambulatory setting, more than half of them had not taken any treatment and less than 20% had taken an antiresorptive drug before hip fracture. This implies that this high-risk group was not sufficiently recognized and targeted for evaluation and treatment, which, indeed, demonstrates dramatic underuse of anti-osteoporotic medication for primary hip fracture prevention.

In our series, the increase in osteoporosis medication prescription after hip fracture was more pronounced than in previous papers [10, 1216]. This increment was, namely, consistent for newer bisphosphonate drugs, especially for the ones requiring longer dosing intervals. The improved treatment rate for secondary prevention may largely rely on the fact that anti-osteoporotic drugs are dispensed without charge by our national health care system, following vertebral and femoral fracture. On the other hand, the increased post-fracture use of drugs in our country could also reflect a more responsible attitude of Italian orthopedic surgeons and primary care physicians [1619], along with better patient education and self-advocacy [20]. Moreover, the prevalence of treatment after fracture in our study was somewhat higher than in previous reports also because of the later time period of our survey, which took place some years after the approval of several anti-osteoporotic drugs in Italy. Bearing in mind all these considerations, it is unwise to make direct comparisons with other published reports, owing to different regulatory systems of national health services or different situations (i.e., non ambulatory patients or different periods of observations).

Despite these partially positive results, there appears to be a lot of room for improvement, since a high percentage of patients in the study group remained untreated after hip fracture. Possible explanations for this gap between evidence-based guidelines and treatment rates have been investigated. Several commonly encountered barriers that obstruct osteoporosis therapy in patients with prior clinical fractures may be linked to physicians, patients, and systems defaults. For instance, the responsibility for care is currently not well defined in the post-fracture setting. In fact, many orthopedists feel they are involved only in the surgical care of the fracture patients, whilst they believe that the major responsibility for the implementation of medical follow-up and treatment of osteoporosis lies with the primary care physicians [17]. On their side, primary care physicians may not recognize fractures as sentinel events requiring treatment [12]. Alternatively, they often believe that the occurrence of a hip fragility fracture represents a too-late-stage of osteoporosis, so that, at that time, it becomes impossible to alter the natural history of the disease [11], whose effects on patients’ disability and quality of life are, thus, neglected. In addition, many primary care physicians remain uncomfortable with osteoporosis management [21], or refuse to treat elderly patients, especially males, or those with dementia [15, 17, 22]. Finally, sometimes, they question the efficacy of osteoporosis medications and are reluctant to modify their practice patterns, despite numerous continuing medical education courses about the effectiveness of several drugs approved by health care authorities. Patient-related barriers include the low priority given to osteoporosis within their general health status, in respect to concerns about comorbidity [23]. Additionally, patients often fail to differentiate osteoporosis from acute fracture, believing that, once the latter is healed, they either do not need treatment for osteoporosis or it is too late for therapy [22]. Systems-related barriers include little communication between orthopedic surgeons and primary care physicians, lack of insurance coverage, and inadequate resources to transport patients for follow-up visits and testing, along with an insufficient number of metabolic bone disease centers [22, 23].

Our data confirm a previous report on the early discontinuation of treatment for osteoporosis [24], but the quitting rate was even higher in our series. After a mean period of 1.4 years, less than half of the patients were continuing some kind of treatment for osteoporosis and more than one out five of them stopped therapy following fracture. Also, in the light of the aforementioned barriers, we investigated various factors which influence the continuation or the discontinuation of medication by applying a logistic regression model to our database. In this way, we obtained some useful information for implementing strategies directed at improving long-term patient compliance. It turned out that women are 50% less likely to stop their medication, while the probability of quitting increases with the aging of a patient or, even more, because of residual pain. These results probably depend on the more severe comorbidity of the male patients and on the possible mood swings with aging or pain persistence. On the other hand, the most relevant finding from our data is that a patient who has not undergone diagnostic tests is more than twice as likely (2.46) to stop medication. This probably simply reflects the motivation of those who had previously performed a bone mass measurement being more concerned about the disease. In this context, it is also important to emphasize that the first period after fracture, especially in the oldest patients, is the most critical one, in that we observed the highest drop off rate. This suggests that the immediate post-fracture period is the most important one; therefore, both a close medical and nursing assistance program is desirable, even in this cohort of a relatively healthy population.

Our study has several important strengths. One of these is that we investigated, nationwide, a very large sample of patients with the same ominous low-energy fracture. Moreover, unlike other reports on this topic [9, 11, 19, 24], this study also included men, who currently represent a significant portion of fractured patients. Furthermore, besides post-fracture variation in treatment rates, we also investigated long-term compliance and its determining factors, which are known to have outstanding importance in the management of every chronic disease. Finally, the reliability of our results is strongly supported by the high level of concordance (95%) between the data obtained from the questionnaires and the information provided by the case records of general practitioners, which were available in a fairly relatively high percentage of cases (n=223) (data not shown).

There are several potential limitations of the study as well. First of all, our results depended on patients’ self-reports. Indeed, some patients may have inaccurately reported their data or may have been hesitant to respond with negative comments because the survey came from their physicians. However, this would actually reinforce the conclusions of the study about the undertreatment of osteoporosis. Furthermore, given that all patients had full drug benefits, this may limit the generalizability of the results to other countries, which have different insurance systems. Again this limitation, which deletes cost barriers, strengthens the main conclusions of the study. Thirdly, the inference from our findings is limited to our particular type of patients (i.e., ambulatory with previous hip fracture); therefore, our results cannot be transferred to those living, for example, in a nursing home. Similarly, different results might be expected from a relatively less healthy population, incapable of filling in the questionnaire. Finally, our questionnaires almost exclusively explored drug therapy, so we did not have information about other possible fracture-preventing measures, such as correction of lifestyle factors, home safety evaluations, lower extremity strengthening, or the use of hip protectors, whose role is largely underestimated by most Italian physicians.

In conclusion, our data confirm that few elderly patients who have experienced a hip fracture are adequately treated to prevent future fractures. In addition to a low rate of primary prevention treatment and still insufficient post-fracture therapy, the picture is compounded by the high rate of early discontinuation of osteoporosis medication.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2005