Pain-related fear, catastrophizing and pain in the recovery from a fracture

1 Introduction

Despite seemingly successful medical treatment, poor recovery and pain after a fracture is surprisingly common. Only 30–75% of patients regain pre-fracture levels of physical function during the coming year and just as many have pain (Kaukonen et al., 1988; ; Mossey et al., 1989;Ponzer et al., 1999). The healing of fractures requires immobilization and after fixation there is a period of regaining strength and mobility in the affected joints. Distal radius fractures of the wrist account for 20% of all fractures (Toh and Jupiter, 1994; O’Neill et al., 2001), but complications occur in more than one-third of the patients (Atkins et al., 1990; Lagerström et al., 1998). Fractures of the ankle account for one-tenth of all fractures and delayed recovery is also frequent (Nilsson, 1969; Jensen et al., 1998). When patients fail to gain full recovery the probability increases that the patient will suffer more pain, frustration and disability. Because fractures have a clear initiation they are excellent for studying the development of long-term problems following an acute injury.

Psychological factors have been linked to the recovery process after a fracture, as well as to the development of persistent pain problems. Anxiety is a common feature of fractures (Kennedy et al., 2004) and it is linked to a host of psychological factors including worry and attention that may influence pain perception (Asmundson et al., 2004). Indeed, anxiety has been directly linked to post-fracture pain levels (Jelicic and Kempen, 1999; Falch et al., 2003).

A cognitive behavioral model of how persistent problems develop is the fear-avoidance model (Vlaeyen and Linton, 2000). The basic tenet is that a vicious circle may develop if the pain results in catastrophizing and fear since these are associated with hypervigilance, and avoidance of movement. While these processes may be adaptive in the acute stage (Leeuw et al., 2007), they paradoxically may prolong recovery. There is an evidence to support the model for chronic low back pain (Vlaeyen and Linton, 2000; Leeuw et al., 2007) but there is a paucity of studies concerning how catastrophizing and fear may affect acute pain. Fractures offer a unique opportunity to study this model.

In a prospective study, fear predicted disability 6 months after an acute back pain episode (Swinkels-Meewisse et al., 2006). Similarly, another study of 44 patients seeking care for acute low back pain showed that an increasing level of fear and catastrophizing during the initial two weeks’ period was related to poorer recovery 12 months later (Sieben et al., 2002). One reason may be that patients who expect a quick recovery but who do not experience it then become more worried (Sieben et al., 2002; Rachman and Arntz, 1991). Thus, pain-related fear and catastrophizing in the acute phase may be important for future pain perception and recovery, but little is known about the process particularly in disorders other than low back pain.

The purpose of this study was to investigate the role of pain-related fear and catastrophizing during the acute phase on long-term recovery from a fracture. We predicted that high or increasing levels of fear and catastrophizing would be related to more pain and less recovery at follow-up.

2 Methods

2.3 Assessments

2.3.1 Pain-related fear and catastrophizing

Fear-avoidance beliefs were measured with a modified form (MFABQ) of the Fear-Avoidance Beliefs Questionnaire (FABQ) which has good reliability and validity (Waddell et al., 1993) and in a pilot test the MFABQ correlated highly with the full FABQ (n = 36; r = 0.97). Four items were selected from the factor that deals with the relationship between physical activity and pain to form the modified version (MFABQ), and which has previously been successfully employed (Linton et al., 2000; Buer and Linton, 2002). Participants were asked to rate items in relation to their own prior experiences of pain and the items concerned the beliefs that:

1. Pain is caused by physical activity;

2. Physical activity makes one’s pain worse;

3. Physical activity might be harmful;

4. One should not do physical activities which (might) make one’s pain worse.

Consistent with the FABQ, the items were answered on a verbal (Likert type) seven-point scale (0–6, sum score 0–24), from “strongly disagree” to “strongly agree”, higher sum score indicate stronger fear-avoidance beliefs.

Catastrophizing was measured by The Pain Catastrophizing Scale (Sullivan et al., 1995; Severeijns et al., 2002) which is a self-report scale that consists of 13 items, with a five-point scale (0–4, sum score 0–52), from “not at all” to “all the time”.

2.3.2 Pain and expected recovery

Self-ratings of current pain and expected recovery were assessed at T1–T3. A Numeric Rating Scale (NRS) 0–10 was used to rate pain, where 0 meant “no pain” and 10 meant “unbearable pain”. Similarly, patients estimated their risk of having problems from the fracture for more than an additional 6 months in percent of recovery compared to pre-fracture status (e.g. 100% recovered or 50% recovered) at T1 and T2 (expected recovery) as well as actual recovery at T3.

Fig. 1

An overview of the percentage of patients rating expected and actual recovery at T1–T3.

2.3.3 Mobility and strength

Mobility of the wrist or ankle and strength was assessed at the 9-month follow-up. Mobility was measured in the affected area (wrist/ankle) using a 310 mm double arm plastic goniometer with a half-circle protractor scale from 0° to 180° with 1° increments (KEBO Care^® , Sweden). This was done in 6 directions, flexion, extension, radialflexion, ulnarflexion, pronation and supination for the wrist and in 2 directions for the ankle, flexion and extension. The reliability of goniometer mobility measures has been debated, but the device is accurate enough to detect changes and differences between left and right sides, which was the purpose of this study (Horger, 1990; Sölveborn and Olerud, 1996).

Handgrip strength was measured using the JAMAR^® hand dynamometer (Sammons^® Preston, Bolingbrook, IL, USA) which registers the maximal momentary strength (Lagerström et al., 1998; Peolsson et al., 2001). It has been found to be a reliable method which is recommended for clinical use (Peolsson et al., 2001). For the patients with a fracture of the ankle, a standardized standing heel-rise test was employed (Svantesson et al., 1998).

3 Statistical analyses

The analyses were made with the SPSS (version 11.0) statistical software program. Because of some skewed distributions, non-parametric statistics were used. Since the hypothesis was that high or increasing levels of fear and catastrophizing would lead to poorer recovery, patients were classified into groups. To obtain a cut-off for categorization, the median (M = 9) on the MFABQ in the general population in Sweden was employed (Buer and Linton, 2002). Participant’s scores at T1 and T2 were cast against this cut-off and categorized on both T1 and T2 and then combined. Consequently, participants could have the same level (consistently high or consistently low) or different levels (increasing, alternatively decreasing) between T1 and T2 resulting in four profiles (High T1–High T2, consistently high; Low T1–Low T2, consistently low; Low T1– High T2, increasing; High T1–Low T2, decreasing). In a further step to consolidate the data for analyses of either those patients having a high score on the MFABQ at both T1 and T2 or who had an increasing level between T1 and T2 were classified as High. Likewise, either those who had a low score at both time points or who had a decrease from high to low were classified as Low. A similar procedure was done for the catastrophizing scores on the PCS. Here the median from the general population was 11 (Buer and Linton, 2002) and similar groups were formed as for fear-avoidance beliefs above. Thus, those with high scores could then be compared with those having low scores and odds ratios calculated concerning the prospective relationship to outcome at T3.

In addition to self-ratings of recovery, patients were also classified as “recovered” or “not recovered” based on the objective assessments of mobility and strength. Since pre-fracture measurements were not possible, we used the intra-individual difference between the injured and the uninjured side as a proxy for comparison. If the ratio between the injured side and the uninjured side was ≤10% the case was classified as Recovered and if the difference was >10% as Not Recovered.

Logistic regression analyses were performed in order to calculate odds ratios which are reported with 90% confidence intervals as an estimate of risk. Because this is an exploratory study examining the possible role of fear and catastrophizing we choose the 90% since the study group is heterogeneous and the sample size is limited.

4 Results

5 Discussion

This is the first investigation to our knowledge of the role of painrelated catastrophizing and fear in the recovery from a fracture. Our results show that higher than average scores, or increasing scores, on pain-related fear and catastrophizing are associated with future pain and recovery levels as measured by questionnaires and muscle strength. Indeed, higher levels increased the risk of having more intense pain by three-fold and the risk of having poorer recovery in terms of muscle strength by six-fold. These findings are in line with other research concerning back pain and support the fear-avoidance model (Leeuw et al., 2007; Vlaeyen et al., 1995; Lohnberg, 2007; Vangronsveld et al., 2007).

This effect may be related to expectations since the vast majority of patients, at the time of injury, expected a full recovery within six months, but less than a third actually achieved this. Indeed, expectations of a full recovery dropped by about half from T1 to T2 and self-reported recovery was slower than anticipated for many. Indeed, participants seemed to adjust their expectations to the reality of the recovery process. This change in expectation might influence fear and catastrophizing but our previous research suggests that fear and catastrophizing may well drive expectations (Boersma and Linton, 2006). A mechanism may be that when expectations are not realized then this results in increases in fear and catastrophizing that in turn hinder recovery. This finding and the interpretation are in line with Sieben (Sieben et al., 2002) where increases in fear in patients seeking care for an acute back pain episode were also related to actual recovery.

On average, fear-avoidance beliefs and catastrophizing showed moderate levels and decreased over time as would be anticipated during healing. However, given the results above, this may create a “false sense” of proper recovery for health care professionals which might lead to a failure to identify patients risking a poor recovery. Our findings suggest that it is important to monitor pain-related fear and catastrophizing during the first weeks of the recovery process. Further, psychologically oriented interventions might then be helpful in order to prevent complications and poor recovery. Because patient expectations are not in line with the usual course of recovery, one method might be to provide specific information tailored to those at risk, e.g. about the natural course of recovery (Linton et al., 2007). Further, since fear and catastrophizing are known to be related to avoidance of movement a program of exposure exercises may be worthwhile (Vlaeyen and Linton, 2000; Leeuw et al., 2007;Lohnberg, 2007).

There are some methodological idiosyncrasies which should be kept in mind when interpreting these results. First, the size of population restricts the analyses that may be conducted and potentially the generalization of the findings. With a very large sample it would have been possible to conduct analyses using specific fracture locations, age groups, gender, and risk profiles. Nevertheless, our sample is sufficient to study the main question, i.e. the role of fear and catastrophizing in recovery. Second, the time frame for measurement might have been different, e.g. with several assessments at shorter intervals which would have allowed for a more detailed analysis.

6 Conclusions and implications

Our results show that fear and catastrophizing may be instrumental in the process of recovery after a fracture and suggest that early identification and preventive interventions might enhance recovery. Future studies will need to replicate, generalize, and test more specific hypotheses.