Hip or spine surgery first?

Abstract

Aims

Patients may present with concurrent symptomatic osteoarthritis (OA) of the hip and degenerative disorders of the lumbar spine, with surgical treatment being indicated for both. Whether arthroplasty of the hip or spinal surgery should be performed first remains uncertain.

Materials and Methods

Clinical scenarios were devised for a survey asking the preferred order of surgery and the rationale for this decision for five fictional patients with both OA of the hip and degenerative lumbar disorders. These were symptomatic OA of the hip and: 1) lumbar spinal stenosis with neurological claudication; 2) lumbar degenerative spondylolisthesis with leg pain; 3) lumbar disc herniation with leg weakness; 4) lumbar scoliosis with back pain; and 5) thoracolumbar disc herniation with myelopathy. This survey was sent to 110 members of The Hip Society and 101 members of the Scoliosis Research Society. The choices of the surgeons were compared among scenarios and between surgical specialties using the chi-squared test. The free-text comments were analyzed using text-mining.

Results

Responses were received from 51 hip surgeons (46%) and 37 spine surgeons (37%). The percentages of hip surgeons recommending ‘hip first’ differed significantly among scenarios: 59% for scenario 1; 73% for scenario 2; 47% for scenario 3; 47% for scenario 4; and 10% for scenario 5 (p < 0.001). The percentages of spine surgeons recommending ‘hip first’ were 49% for scenario 1; 70% for scenario 2; 19% for scenario 3; 78% for scenario 4; and 0% for scenario 5. There were significant differences between the groups for scenarios 3 (more hip surgeons recommended ‘hip first’; p = 0.012) and 4 (more hip surgeons recommended ‘spine first’; p = 0.006).

Conclusion

In patients with coexistent OA of the hip and degenerative disorders of the spine, the question of ‘hip or spinal surgery first’ elicits relatively consistent answers in some clinical scenarios, but remains controversial in others, even for experienced surgeons. The nature of neurological symptoms can influence surgeons’ decision-making.

Cite this article: Bone Joint J 2019;101-B(6 Supple B):37–44.

Patients with symptomatic osteoarthritis (OA) of the hip often have concomitant lumbar or thoracolumbar spinal disorders.^1,2 According to the Medicare claims database, 4.5% of patients who undergo total hip arthroplasty (THA) have had lumbar surgery within five years of the hip surgery.² Usually in patients with such symptoms from the hip and spine, the priority of treatment is determined by the severity and location of symptoms, activities of daily living, and the preference of the patient. However, when the symptoms from both are severe enough to warrant surgery, it may be difficult to decide on the optimal order of treatment. There are more complications, such as dislocation, and less satisfaction after THA in patients with lumbar symptoms or previous lumbar fusion.^2,3 One study, however, suggested that patients with lumbar symptoms and coexistent OA of the hip have some relief from their back symptoms after THA and may not then require spinal surgery.⁴

One study suggested a treatment pathway for patients with concurrent hip and spinal disorders based on a single institutional protocol.¹ This pathway of treatment was not evidence-based, did not evaluate many characteristics of the different concomitant disorders, and the recommendations for several typical scenarios were incomplete. For example, there is general consensus that a patient with OA of the hip and a ‘progressive neurological deficit’ warrants an ‘urgent spinal opinion’. However, the order of treatment is not obvious if the neurological deficit is chronic and non-progressive, such as in chronic lumbar radiculopathy or stenosis with weakness. When either the hip or spinal symptoms are severe and predominant, the decision about the order of treatment may be straightforward. However, if the symptoms from both regions are equally severe, or if the radiological or pathological aspects of one disorder appear to influence the surgical treatment of the other, the decision for the specific sequence of treatment may be complex. Preoperative counselling would be easier if there was consensus between hip and spine surgeons on the optimal order of treatment.

Therefore, a study involving a survey with five clinical scenarios of concurrent OA of the hip and common lumbar spinal disorders was designed by a senior arthroplasty and spine surgeon and sent to the members of two specialist orthopaedic societies to elicit opinions about the order and rationale of treatment. The primary aim was to determine the preference and rationale for the order of treatment for patients with OA of the hip and five different lumbar disorders. The hypothesis was that surgeons specializing in THA would have different preferences for the sequence of treatment than surgeons specializing in spinal surgery.

Materials and Methods

A survey including five fictional clinical scenarios of patients with concomitant OA of the hip and common lumbar and thoracolumbar spinal disorders was designed. The patients had failed extensive nonoperative treatments and were sufficiently disabled to warrant surgery for both disorders. Each scenario provided detailed clinical history, physical examination, plain radiographs, and MR imaging of the spine (Fig. 1). The socioeconomic status of the patient and their preferences for treatment were not included. The scenarios involved patients with painful OA of one hip and: 1) degenerative lumbar spinal stenosis with neurogenic claudication; 2) degenerative lumbar spondylolisthesis with leg pain; 3) a single-level lumbar disc herniation with weakness in the leg; 4) lumbar scoliosis with sagittal imbalance and back pain; and 5) thoracolumbar disc herniation with signs of myelopathy. The clinical information and imaging studies of the scenarios are shown in Supplementary Figure a.

Fig. 1.

A computer screenshot showing the layout of the survey. Only part of scenario 4 is shown. See Supplementary Figure a for the full survey.

The participants were asked to give their choice of the order of treatment (THA first or spinal surgery first or ‘no preference’) and their rationale. The arthroplasty surgeons were also asked for their preference of the choice of articulation (< 32 mm femoral head, > 36 mm femoral head, dual-mobility component, constrained component, or no preference), if they chose to perform THA first. The survey was completed using SurveyMonkey (SurveyMonkey, Menlo Park, California).

The eligibility criteria for the participants were: surgeons who undertook THA and were members of the North American Hip Society,⁵ and spine surgeons who were members of the Scoliosis Research Society (SRS).⁶ The arthroplasty and spine surgeons were enrolled separately. A letter from a past president of The Hip Society (PFL) requesting participation with the survey was sent by The Hip Society coordinator, on two different days, by email to the 110 members. Many of these surgeons only perform hip preservation surgery or arthroscopy of the hip and thus declined participation. For enrolment of the spine surgeons, an author (KBW) selected participants from 101 members of the SRS based on the pattern of their practice or attendance at meetings, and they were contacted individually by alphabetical order. The latter approach is based on the availability sampling method,⁷ and has been previously used in orthopaedic research to ensure the qualification of respondents.⁸

The choices of the order of treatment for the five scenarios were compared between the two surgeon specialists (THA vs spinal). The free-text comments of the surgeons were analyzed by two methods to determine the rationale for the decision in each scenario. Text-mining was initially used to identify the most frequently used words in each scenario.⁹ Second, an author (KBW) read the comments and summarized the key points. Thus, the most frequently used words in text-mining could be explained by the key points summarized by a different author. For example, ‘recovery’ was a frequently used word of the THA surgeons in scenario 1; from the key points the author summarized, this was because many believe that untreated neurogenic claudication can impede the recovery from THA.

Statistical analysis

Descriptive statistics of the responders’ demographics and rates of response were reported in mean, median, and percentage. The chi-squared test (or Fisher’s exact test when the expected count was fewer than five) was used to compare the choices between the surgeon specialists (THA vs spinal). The frequency with which words were used was calculated as the number of specific words divided by the total word count.⁹ R version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria) was used for data analysis and text-mining. The α value was set at 0.05.

Results

Responses to the survey were received from 88 surgeons, with a response rate of 46% (51/110) for the THA surgeons and 37% (37/101) for the spine surgeons. The mean years of experience after completion of training was 30.8 (14 to 60) for the THA surgeons and 23.4 (5 to 34) for the spine surgeons (Fig. 2). The widespread geographical locations of the surgeons are shown in Fig. 3.

Fig. 2.

Graphs showing the years of experience after surgical training for the respondents for: a) hip arthroplasty; and b) spine.

Fig. 3.

Geographical location of the respondents.

The percentages of THA surgeons recommending ‘hip first’ differed significantly among the five scenarios (χ² = 44.5; p < 0.001; Table I). They were more likely to choose ‘hip first’ in scenario 2 (OA of the hip and spondylolisthesis and leg pain, 73%), and less likely to choose ‘hip first’ in scenario 5 (OA of the hip and thoracolumbar disc herniation with myelopathy, 10%).

Table I.

The choices for the order of treatment of the surgeons in the five scenarios. The p-values were based on comparing the percentage of choosing ‘hip first’ between the two specialties

Scenario	Spine first, n (%)	Hip first, n (%)	No preference, n (%)	Spine vs hip arthroplasty, χ-value; p-value*	Pattern noted
1†				0.532; 0.466	Choices were variable within each of the two specialties.
THA surgeon	17 (33)	30 (59)	4 (8)
Spine surgeon	17 (46)	18 (49)	2 (5)
2‡				0.001; 1.000	Choices were relatively consistent between specialties, both preferring ‘hip first’.
THA surgeon	9 (18)	37 (73)	5 (10)
Spine surgeon	3 (8)	26 (70)	8 (22)
3§				6.259; 0.012¶	Choices were different between specialties, more spine surgeons preferring ‘spine first’.
THA surgeon	23 (45)	24 (47)	4 (8)
Spine surgeon	27 (73)	7 (19)	3 (8)
4**				7.522; 0.006¶	Choices were different between specialties, fewer spine surgeons preferring ‘spine first’.
THA surgeon	24 (47)	24 (47)	3 (6)
Spine surgeon	4 (11)	29 (78)	4 (11)
5††				2.234; 0.071	Choices were relatively consistent between specialties, both preferring ‘spine first’.
THA surgeon	44 (86)	5 (10)	2 (4)
Spine surgeon	36 (97)	0 (0)	1 (3)

^*Chi-squared test

^†Lumbar canal stenosis with neurogenic claudication combined with osteoarthritis of the hip

^‡Degenerative lumbar spondylolisthesis with radicular leg pain combined with osteoarthritis of the hip

^§Lumbar disc herniation with muscle strength weakness combined with osteoarthritis of the hip

^¶Statistically significant

^**Scoliosis with back pain combined with osteoarthritis of the hip

^††Thoracolumbar disc herniation with myelopathy combined with osteoarthritis of the hip

The choices of the order of treatment were compared between the specialties (THA vs spine; Table I and Fig. 4). Overall, the choices were relatively consistent within each specialty and between the two specialties for scenario 2 (OA of the hip and spondylolisthesis and leg pain), with surgeons in both groups preferring THA first, and scenario 5 (OA of the hip and thoracolumbar disc herniation with signs of myelopathy), with surgeons in both groups preferring spinal surgery first. There was a significant difference between the groups for scenario 3 (OA of the hip and lumbar disc herniation with weakness), with 19% of the spine surgeons preferring THA first, and 47% of THA surgeons preferring THA first (p = 0.012). There was also a significant difference between the groups for scenario 4 (OA of the hip and lumbar scoliosis with back pain), with 78% of spine surgeons preferring THA first, and 47% of THA surgeons preferring THA first (p = 0.006). For scenario 1 (OA of the hip and lumbar stenosis with neurogenic claudication), the choices were variable within each specialty: approximately half of the spine surgeons and 59% of THA surgeons preferred THA first.

Fig. 4.

Graphs showing the percentage of surgeons choosing ‘hip surgery first’ in the five scenarios: a) scenario 1, degenerative lumbar spinal stenosis with neurogenic claudication; b) scenario 2, degenerative lumbar spondylolisthesis with leg pain; c) scenario 3, a single-level lumbar disc herniation with weakness in the leg; d) scenario 4, lumbar scoliosis with sagittal imbalance and back pain; and e) scenario 5, thoracolumbar disc herniation with signs of myelopathy.

Three patterns of preference for the order of treatment were noted. First, weakness in the leg related to a herniated disc (scenario 3) was of greater concern to spine surgeons than THA surgeons, as more spine surgeons chose ‘spine first’, but radicular leg pain due to stenosis and spondylolisthesis (scenario 2) was of less concern to both groups. Second, in the setting of scoliosis and coexistent OA of the hip (scenario 4), most spine surgeons were reluctant to perform spinal surgery first but approximately half of the THA surgeons preferred THA first. Third, treating the spine first was preferred by most surgeons in both groups when there were signs of myelopathy (scenario 5).

The rationale for the order of treatment was evaluated from the most frequently used words in the surgeons’ comments, as identified by text-mining (italicized words) and plotted for each scenario (Fig. 5).

Fig. 5.

Graphs showing the most frequently used words in the surgeons’ comments in the five scenarios: a) scenario 1, degenerative lumbar spinal stenosis with neurogenic claudication; b) scenario 2, degenerative lumbar spondylolisthesis with leg pain; c) scenario 3, a single-level lumbar disc herniation with weakness in the leg; d) scenario 4, lumbar scoliosis with sagittal imbalance and back pain; and e) scenario 5, thoracolumbar disc herniation with signs of myelopathy. OA, osteoarthritis; VAS, visual analogue scale.

Comments in scenario 1 (OA of the hip and degenerative lumbar spinal stenosis with neurogenic claudication) included: decision was dependent on which problem the surgeon (either THA or spinal) perceived as more ‘severe’, ‘dramatic’, ‘symptomatic’, or ‘worse’. Many THA surgeons commented that THA can improve spinal-pelvic biomechanics and lessen spinal symptoms and should be performed first. Surgeons in both groups commented, however, that untreated neurogenic claudication could impede the ‘recovery’ from THA and recommend treating the spine first.

Comments in scenario 2 (OA of the hip and degenerative spondylolisthesis with leg pain) included: many surgeons in both groups commented that, without a neurological motor deficit, the hip problem was more ‘severe’ or ‘advanced’ than radicular leg pain, and should be treated first by THA, which was considered to provide more predictable pain ‘relief’ than spinal ‘fusion’ and thus was preferred first.

Comments in scenario 3 (OA of the hip and single level lumbar disc herniation with weakness) included: many surgeons in both groups commented that muscular ‘weakness’ was an important ‘neurological deficit’ and ‘discectomy’ was relatively straightforward for an extruded ‘disc’, and thus preferred spinal surgery first.

Comments in scenario 4 (OA of hip and adult scoliosis with sagittal imbalance and back pain) included: many THA surgeons commented that balancing the spine first was important for optimizing the ‘position’ of the components at subsequent THA, because this allowed the ‘change’ in spinopelvic alignment due to the spinal surgery to be considered at the time of THA. However, many THA surgeons would still treat the hip first, as the patient was ‘neurologically’ intact and the spinal surgery was not urgent. For many spine surgeons, the preference to treat the hip first was a practical consideration, as THA, compared with surgery for scoliosis, was an ‘easier’ procedure with more ‘predictable’ outcomes and a quicker ‘recovery’.

Comments in scenario 5 (OA of hip and thoracolumbar disc herniation with signs of myelopathy) included: ‘myelopathy’ due to spinal ‘cord compression’ was considered to be more urgent than OA of the hip.

For the choice of THA articulation, most of the 51 hip surgeons chose a large head (36 mm, 40 mm, or larger) for all five scenarios (Fig. 6). A dual-mobility component was most likely to be chosen for scenario 4 (12 the THA surgeons (24%)) than for scenarios 1 (12%), 2 (16%), 3 (8%), and 5 (10%), but this difference was not statistically significant (χ² = 6.6; p = 0.16). No surgeon chose a constrained component for any scenario.

Fig. 6.

Articulation choices by 51 hip arthroplasty surgeons. No surgeon chose a constrained component for any scenario.

Discussion

To our knowledge, this is the first survey of preferences for the order of treatment for patients with OA of the hip and concomitant lumbar spinal disorders. The results showed some patterns of preference shared by both specialties in some scenarios, such as treatment of the hip first when the spinal problem was radicular leg pain and treatment of the spine first with signs of myelopathy, but significantly varied opinions in others, such as weakness in the leg due to a single-level lumbar disc herniation and mechanical back pain from scoliosis. In both groups, the nature of neurological deficits influenced the thinking, with the most time-sensitive issues such as myelopathy and weakness driving the decision-making towards performing spinal surgery first.

Previous studies have suggested complex interactions between the two pathologies.^2,10-12 While the presence of lumbar pathology can adversely affect the outcome of THA,^2,10 whether a successful THA may ease spinal symptoms remains controversial. In a prospective study of 25 patients with OA of the hip and lumbar spinal symptoms, THA reduced the intensity of low back pain and the Oswestry disability score¹³ by 35% and 34%, respectively.¹¹ However, a study based on the Swedish national registries reported that a previous THA adversely affected the outcome of lumbar surgery, and at one year of follow-up the mean visual analogue scale for back pain in patients with a previous THA was significantly higher than that in patients without a previous THA, probably due to the delay in spinal surgery or a reduced ability to participate in spinal rehabilitation.¹² Thus, further investigation of the optimal order of treatment in these patients is required.

From the comments in this study, the nature and severity of neurological symptoms appears to be an important factor in the decision-making process. Weakness in the leg (scenario 3) was of greater concern to spine surgeons, as more preferred ‘spine first’, than THA surgeons. However, radicular leg pain only (scenario 2), although a neurological finding, was considered of lesser concern to both groups, as < 20% of surgeons in both specialties selected spinal surgery first. Neurogenic claudication (scenario 1), when compared with the above two symptoms, was considered as ‘medium’ for urgency of treatment. This could be explained by the belief that motor loss, as compared with radicular pain or neurogenic claudication, is usually a more time-sensitive issue. Unlike radicular pain and neurogenic claudication, which are often reversible in nature, the presence of motor loss may lead to permanent disability without timely treatment.

We did not explore the complex issue of THA in patients with spinal imbalance, which has many manifestations and treatment algorithms. However, our results suggested near consensus among spine surgeons in favour of ‘hip surgery first’ in these patients. A balanced spine is defined as one in which the head and neck are positioned squarely over the sacrum and pelvis; the vertical axis of the head should fall in the centre of the sacrum and the midpoint of a line connecting the two femoral heads.¹⁴ Under physiological conditions, this balance is maintained by the lordotic curvatures of the cervical and lumbar spines and, in between, the kyphotic curvature of the thoracic spine. Degenerative OA of the hip has been found to increase lumbar lordosis in some patients. This may aggravate spinal stenosis, but it rarely affects the coronal curvature of the spine.¹⁵ Even so, the responses to scenario 4 showed that most spine surgeons were reluctant to proceed first. Their overall rationale was somewhat indirect as it was more related to the benefits of hip surgery first than concern about spinal surgery first. Future studies with greater sample size would help to reveal the underlying reason. Our results may aid in advising patients when discussing a referral to a spine colleague.

The choice of the articulation at THA in this study may indicate the usual preference of THA surgeons or a greater concern about dislocation in a patient with a coexistent spinal problem. The survey did not ask the THA surgeon for their usual choice of articulation. A component with a large head was chosen by most surgeons for all scenarios. For scenario 4, in which there may be a higher risk of dislocation, 24% of the THA surgeons chose a dual-mobility component, which may reduce the risk of dislocation in both primary and revision surgery.¹⁶ Additional studies of THA in these patients are needed to determine the optimal choice for the size and design of the articulation.

It is important to note that the present study was not to put the opinion of the surgeon at the centre of clinical decision-making. Instead, the findings should be used to inform patient-centred shared decision-making, which is the benchmark for healthcare delivery. It is the duty of surgeons not only to elicit as much information from patients and evaluate every nuance of symptoms between the two pathologies, but also to confer on patients as much knowledge as possible about their condition to allow an informed preference to be made. Also, surgeons should inform patients about the relative risks, recovery period, and outcomes after discussing each surgical option.

This study has limitations. The response rate was 46% for the members of The Hip Society and 37% for the members of the SRS, which may limit the generalization of the results to some patterns of practice. The socioeconomic status of the patient and their preferences for treatment were not considered. Another major limitation is that the dynamic biomechanical relationships between the movements of the hip, pelvis, and spine, in which the pathophysiology of the spine and pelvis influence the outcome of THA, were not considered. As a survey, this study provides a snapshot of current opinion in North America regarding this topic. Prospective studies, including observational studies of the natural history and interventional studies comparing outcomes between different orders of treatment are needed to develop patient-centred and evidence-based treatment protocols for patients with this dual pathology. Our study serves as a practical starting point for such future efforts.

In conclusion, with concurrent degenerative disorders of the hip and spine, the question of hip or lumbar surgery first elicits relatively consistent answers in some clinical scenarios, but remains controversial in others, even for experienced hip and spine surgeons. The nature of neurological symptoms is an important determinant in surgeons’ decision-making. Our findings may serve as a reference facilitating meaningful discussion and decision-making for these patients. The findings can also be used to assist the generation of hypotheses and study design for future investigations in this field.

Take home message

- In patients with coexistent osteoarthritis of the hip and degenerative spinal disorders, there remains uncertainty about which procedure, total hip arthroplasty or spinal surgery, should be performed first, when surgical treatment is indicated for both disorders.

- The optimal order of treatment is relatively predictable in certain clinical scenarios, but remains controversial in others, even for experienced surgeons, with the decision influenced by the nature of neurological symptoms.

- Myelopathy, weakness, neurogenic claudication, and radicular pain, in that order, drive the decision-making towards performing spine surgery first in this dual-pathology setting.

Twitter

Follow the authors @StanfordMed, @StanfordHealth, and @DukeHealth