DBS indicates deep brain stimulation; PD, Parkinson disease.
DeLong MR, Huang KT, Gallis J, Lokhnygina Y, Parente B, Hickey P, Turner DA, Lad SP. Effect of Advancing Age on Outcomes of Deep Brain Stimulation for Parkinson Disease. JAMA Neurol. 2014;71(10):1290-1295. doi:10.1001/jamaneurol.2014.1272
Deep brain stimulation (DBS) is a well-established modality for the treatment of advanced Parkinson disease (PD). Recent studies have found DBS plus best medical therapy to be superior to best medical therapy alone for patients with PD and early motor complications. Although no specific age cutoff has been defined, most clinical studies have excluded patients older than 75 years of age. We hypothesize that increasing age would be associated with an increased number of postoperative complications.
To evaluate the stepwise effect of increasing age (in 5-year epochs) on short-term complications following DBS surgery.
Design, Setting, and Participants
A large, retrospective cohort study was performed using the Thomson Reuters MarketScan national database that examined 1757 patients who underwent DBS for PD during the period from 2000 to 2009.
Main Outcomes and Measures
Primary measures examined included hospital length of stay and aggregate and individual complications within 90 days following surgery. Multivariate logistic regression analysis was used to calculate complication-related odds ratios (ORs) for each 5-year age epoch after controlling for covariates.
Overall, 132 of 1757 patients (7.5%) experienced at least 1 complication within 90 days, including wound infections (3.6%), pneumonia (2.3%), hemorrhage or hematoma (1.4%), or pulmonary embolism (0.6%). After adjusting for covariates, we found that increasing age (ranging from <50 to 90 years of age) did not significantly affect overall 90-day complication rates (OR, 1.10 per 5-year increase [95% CI, 0.96-1.25]; P = .17). The 2 most common procedure-related complications, hemorrhage (OR, 0.82 [95% CI, 0.63-1.07]; P = .14) and infection (OR, 1.04 [95% CI, 0.87-1.24]; P = .69), did not significantly increase with age.
Conclusions and Relevance
Older patients with PD (>75 years) who were selected to undergo DBS surgery showed a similar 90-day complication risk (including postoperative hemorrhage or infection) compared with younger counterparts. Our findings suggest that age alone should not be a primary exclusion factor for determining candidacy for DBS. Instead, a clear focus on patients with medication-refractory and difficult to control on-off fluctuations with preserved cognition, regardless of age, may allow for an expansion of the traditional therapeutic window.
Parkinson disease (PD) affects an estimated 1 million patients in the United States and is a cause of significant disability for those affected.1 For patients with advanced PD who have severe levodopa-induced motor fluctuations and dyskinesia, deep brain stimulation (DBS) has been found to be an effective treatment in reducing motor disability and improving quality of life.2 As with any surgical procedure, DBS has inherent risks and has been associated with postoperative complications, including hemorrhage, wound infection, and pneumonia.3 Recent results from the Controlled Trial of Deep Brain Stimulation in Early Patients with Parkinson’s Disease4 (mean age, 52 years) demonstrated that DBS plus best medical therapy was superior to best medical therapy alone for patients with PD and early motor complications, potentially expanding the therapeutic window to treat patients earlier in the disease process. However, PD remains a chronic progressive disease of elderly populations, with a mean age at diagnosis of approximately 60 years and a disease duration of 14 years before DBS is performed.5,6 The effect of increasing age on DBS candidacy and surgical complications is of increasing importance because the number of individuals with PD is expected to double by 2030.7 The purpose of our study was to describe the population-level risk of adverse outcomes among older patients undergoing DBS for PD and to assess the effect of advancing chronologic age on postoperative complications. We hypothesize that increasing age would result in a higher number of postoperative complications, including risk of hemorrhage, wound infection, pneumonia, and pulmonary embolism in older patients with PD. Critical analyses are needed to provide much needed clinical evidence, influence future treatment strategies, and guide future health care policy.
Data for the study were obtained using the Thomson Reuters MarketScan national database, which contains information from Commercial Claims and Encounters, the Medicare Supplemental and Coordination of Benefits, and the Medicaid databases. Within this database, a cross section of inpatient admissions, inpatient services, outpatient services, outpatient pharmaceuticals, and enrollment tables in the United States between 2000 and 2009 were examined. This database subset was compared with the total number of new DBS implants performed in the United States, as recorded by the manufacturer (Medtronic), and represents approximately 7% of all new DBS implant cases in the same time period. The Duke University institutional review board reviewed and approved this study. Data were obtained from a third-party, de-identified database. This was a large, retrospective study without direct patient contact, required consent, or financial compensation.
Diagnosis (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM]) and procedure (Current Procedural Terminology [CPT]) codes were used to identify patients with paralysis agitans (ICD-9-CM code 332.0) who have undergone an implant of intracranial neurostimulation leads (ICD-9-CM code 02.93 and CPT-4 code 61683 or 61687) for inclusion in our study. Using these criteria, we identified a total of 1757 patients as the study population, including 590 with unilateral implants, 551 with bilateral implants, 122 with staged implants, and the remainder with indeterminate implants.
For purposes of describing the population, patients (at the time of the DBS procedure) were divided into 5-year epochs (ranging from <50 to 90 years of age). Primary outcomes included hospital length of stay and overall complication rate. In addition, specific complication rates for pneumonia, postoperative infection, pulmonary embolism, and intracranial hemorrhage within 90 days were assessed as primary outcomes. Mortality, lead removal or revision, and generator removal or revision within 90 days were enumerated but not analyzed in the statistical model owing to the infrequency of these events.
The effect of age on postoperative complications was assessed using unadjusted and adjusted logistic regression models. In addition, the effect of age on length of stay (dichotomized as 1 day or >1 day) was analyzed using logistic regression. Age was analyzed as a continuous variable, with odds ratio (OR) results given for a 5-year linear increase in age. Multivariate models were adjusted for Charlson Comorbidity Index (dichotomized as 0 or >0), insurance type (Medicare, Medicaid, or commercial), and sex. Insurance status was included in the multivariate model to ensure that age was the variable being assessed, and its removal did not substantially change the results. Several knot points were examined to include a spline term for age in the model, in order to capture a possible nonlinear effect of increasing age, but these spline terms did not significantly change the model. Statistical significance was defined by P < .05. All analyses were conducted using SAS version 9.3 (SAS Institute Inc).
A total of 1757 patients met the inclusion criteria, having undergone DBS for PD between 2000 and 2009, and were included in the analysis. Demographic characteristics of the cohort are listed in Table 1. The mean (SD) patient age was 61.2 (10.1) years, with 582 (33.1%) individuals 65 years of age or older and 123 (7.0%) individuals 75 years of age or older. Of the total study population, 1198 (68.2%) were men, and 642 (36.5%) had a Charlson Comorbidity Index of 1 or greater. The majority of patients had either commercial (940 patients [53.5%]) or Medicare (746 patients [42.5%]) insurance. In total, 132 patients (7.5%) had at least 1 postoperative complication within 90 days of surgery, as defined by the limited subset of ICD-9-CM codes that were analyzed. The most common complication observed within 90 days was wound infection (64 patients [3.6%]), followed by pneumonia (41 patients [2.3%]), hemorrhage or hematoma (25 patients [1.4%]), and pulmonary embolism (10 patients [0.6%]). Within 90 days, lead replacement or revision was performed for 30 patients (1.7%), generator revision or removal was necessary for 21 patients (1.2%), and 6 patients (0.3%) died.
After controlling for sex, Charlson Comorbidity Index, and insurance type, overall perioperative complications within 30 days of surgery were not found to be significantly associated with patient age (per 5-year increment: OR, 1.00 [95% CI, 0.83-1.19]; P=.96) (Table 2). In addition, per 5-year increments, there was no detectable significant association between age and the 30-day incidence of hemorrhagic complications (OR, 0.82 [95% CI, 0.60-1.11]; P = .19), infectious complications (OR, 0.95 [95% CI, 0.71-1.27]; P = .72), or pneumonia (OR, 1.19 [95% CI, 0.87-1.63]; P = .28). Pulmonary embolism and mortality events were sufficiently rare that logistic regression analysis was not meaningful.
Of the 123 patients older than 75 years of age, 9 (7.3%) had a complication within 90 days, as opposed to 123 of the 1634 patients (7.5%) from the younger cohort (P = .93; Table 3). In multivariate logistic regression, per 5-year increment, increasing age was not a significant predictor of having at least 1 complication within 90 days (OR, 1.10 [95% CI, 0.96-1.25]; P = .17). Moreover, per 5-year increments, when individual complications were examined, no direct relationship was observed between age and the risk for intracranial hemorrhage (OR, 0.82 [95% CI, 0.63-1.07]; P = .14), infection (OR, 1.04 [95% CI, 0.87-1.24]; P = .69), pulmonary embolism (OR, 1.53 [95% CI, 0.92-2.54]; P = .10), or pneumonia (OR, 1.28 [95% CI, 0.99-1.64]; P = .06). Finally, per 5-year increment, increasing age was not found to significantly affect length of stay at the hospital following surgery (OR, 0.93 [95% CI, 0.73-1.2]; P = .57), although more than 98.5% of the population had only a 1-day length of stay.
For patients with advanced PD and complications of medication or tremor, multiple prospective randomized clinical trials have demonstrated the utility of DBS compared with pharmacotherapy alone.2,4,8 However, DBS has been associated with surgical complications, including wound infection, pneumonia, hemorrhage, pulmonary embolism, neurologic sequelae, and even death.3,9 It has not been extensively examined whether and how age affects these complications with regard to PD, but it is of great relevance given the association of PD with elderly patients and progressive natural history. Likewise, multiple other factors, including age at PD onset, duration of PD, type of PD and associated symptoms (ie, akinetic/rigid vs tremor predominance), presence of even mild cognitive dysfunction, and nonmotor symptoms, all alter the likely response to DBS surgery. In particular, elderly patients with PD are more likely to present with postural instability and gait difficulty, and they often have a more rapid disease course.10,11
However, the proportion of patients undergoing DBS for PD who are older than 75 years of age has remained relatively stagnant for the past decade, suggesting that neurosurgeons are not actively expanding the therapeutic window for surgical intervention to include the elderly population (Figure). In fact, in a meta-analysis of 22 uncontrolled studies,12- 14 the mean age at surgery across different centers was 58.6 years, despite evidence of similar improvement in motor fluctuations and dyskinesia in elderly individuals. This age imbalance may be due to a concern for serious surgical complications in elderly patients or a dearth of optimal candidates in this age group. Older adults may have greater difficulty recovering from surgery because of decreased physiologic reserve, but many case series15- 17 have reported acceptable rates of adverse outcomes or no increase in morbidity in older patients after a variety of neurosurgical procedures. In addition, there is evidence that elderly patients have a higher rate of deterioration in axial symptoms such as freezing of gait and postural instability following surgery.13,14,18 In general, these symptoms are resistant to both dopaminergic therapy and DBS, may be related to a more rapid disease progression, and negatively affect quality of life.19,20 However, careful preoperative evaluation focusing on axial scores off medication, particularly for gait, may predict better surgical outcome.14 For clinicians, patients, and families considering DBS, informed decision making is challenging because of limited data regarding the risks of adverse perioperative events and postoperative complications associated with advancing age. In counseling patients regarding risk, some rely on the extent of comorbid illness and “physiologic age” rather than age measured in years, whereas others may have strict age cutoffs (ie, >75 years). However, a nationally reported risk of adverse outcomes, as it stands today in this growing and potentially higher-risk population, has yet to be addressed because most series and trials are based in medical centers.
Although few prior studies exist on the subject, Voges et al3 found that the rate of secondary adverse events (including cardiovascular risk factors) increased preferentially for older patients and for those receiving DBS for PD. Yet the relationship between patient age and primary surgical complication rates specific to the PD population was not directly addressed. Because neurosurgeons may tend to avoid operating on elderly patients with PD owing to concern for serious surgical complications, it is important to quantify the risks of DBS surgery as they relate to age. Our analysis revealed no increase in hospital length of stay or rates of aggregate complications with increasing patient age in this selected group, nor were there increases in the rates for wound infection, hemorrhage, pulmonary embolism, or pneumonia. Thus, older patients with PD who were selected to undergo DBS were not at an apparent increased substantial risk for these procedure-specific complications. In addition, preliminary data from 161 patients at our institution suggest a mean persistence of use of 9.4 years with no difference observed between patients older than 75 years of age and their younger counterparts. Paralleling this finding, the mortality rate observed in the MarketScan database was sufficiently low that we feel that surgical intervention can be justified for older patients with an expectation of multiple years of benefit to quality of life.
It should be noted, however, that, per 5-year increment, the risk of pneumonia for our patients bordered on significant (OR, 1.28 [95% CI, 1.00-1.64]; P = .06) and that before we adjusted for covariates, per 5-year increment, the 90-day rate for pneumonia was significantly correlated with age (OR, 1.45 [95% CI, 1.22-1.72]; P < .001). This was not unexpected because the finding that the rates of pneumonia increase for elderly patients is prevalent across many surgical fields. This trend has been established for all noncardiac surgical procedures21 and for thoracic surgery specifically.22 However, because DBS involves only a short period of general anesthesia, it should be more similar to a laminectomy, which demonstrates an unchanged rate of complications with age.23
Health care providers (ie, physicians, neurosurgeons, and nurses) counsel patients and families considering operative procedures regarding the risks and benefits of those interventions. Because many of these risks are not well quantified, these providers often rely on a patient’s comorbid disease burden and ill-defined “physiologic age,” rather than actual chronologic age, as predictors of postoperative adverse outcomes. Case series and this type of large registry approach may represent both selection and publication bias because the actual criteria for selecting patients for surgery is unknown.
Compared with the available literature, our study agrees with prior estimates for DBS-related hemorrhagic and infectious complications, providing support for the validity of our findings (Table 4).2- 4,8,24- 32 When examined individually, the infection rates reported in prior studies appears to vary. Many of the previous studies did not distinguish between superficial skin infections and hardware or respiratory infections, so these complications have been combined for this tabulation. In addition, many of these studies examined complications within different postoperative periods but did not necessarily restrict data collection to (within) 30 days or 90 days after surgery. Specifically, the only other large cohort study (Voges et al3) was restricted to 30 days postoperatively, which likely explains the low infection rates, especially given our observation that infectious complications occur relatively constantly for the first few months postoperatively. Lastly, many of these investigations were of small sample size and not focused on age specifically, and the variability of infectious complication rates may reflect a limited population size or differing age ranges. Ours study has evaluated a nationally selected cohort of patients, reducing the discrepancies and biases that result from selecting patients from particular institutions or states. In addition, because the MarketScan database is comprehensive, it more accurately reflects the variety of practices across the United States. To the best of our knowledge, the present study represents the largest study to date on this subject, and the hemorrhagic and infectious complication rates reported are sufficiently consistent to support the validity and utility of our findings.
Despite these strengths, several limitations are present in our study. Although such a trial specifically examining the effect of age (ranging from <50 to 90 years of age) will likely not be conducted in the near future, this large database analysis is retrospective and nonrandomized. As a result, the findings of our study are susceptible to procedure bias, with healthier aged individuals being preferentially selected for surgical intervention. Also, because patients and procedures were selected using diagnosis and procedure codes, miscoding may be present. We have attempted to reduce this inaccuracy by using both ICD-9-CM and CPT codes, allowing for greater data accuracy. In PD, there are many clinical variables to consider that could not be discerned in our study, including the Unified Parkinson’s Disease Rating Scale, symptom severity, medication usage requirement, and neurocognitive assessments of the elderly. In addition, age at PD onset, duration of PD, type of PD, type of symptoms (tremor vs akinetic/rigid), dyskinesias, other medication complications, and amount of brain atrophy make surgical evaluation of DBS for PD different from many other surgical procedures.
The results of our study may serve to inform multidisciplinary teams of the likely complications older DBS patients may experience. The present study should be helpful in guiding this decision with patients and families and preparing surgical teams for preventing the most likely complications in older patients with PD. However, the ultimate process of screening, counseling, and selecting patients with PD and recommending DBS is clearly multifactorial and personalized for each patient.
Among patients with PD who are older than 75 years of age, the overall complication risk, as well as the risk of postoperative hemorrhage, pneumonia, pulmonary embolism, or infection, remains relatively stable, despite increasing age. This suggests a possible expansion of the therapeutic window traditionally considered for DBS candidates, or at least the removal of age as a rigid exclusion criterion.
Accepted for Publication: April 21, 2014.
Corresponding Author: Shivanand P. Lad, MD, PhD, Division of Neurosurgery, Department of Surgery, Duke University Medical Center, PO Box 3807, Durham, NC 27710 (firstname.lastname@example.org).
Published Online: August 25, 2014. doi:10.1001/jamaneurol.2014.1272.
Author Contributions: Dr Lad had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: DeLong, Huang, Gallis, Turner, Lad.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: DeLong, Huang, Turner, Lad.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Huang, Gallis, Lokhnygina, Lad.
Obtained funding: Lad.
Administrative, technical, or material support: Huang, Parente, Lad.
Study supervision: Turner, Lad.
Conflict of Interest Disclosures: Dr Lad has consulted for and received grant support from Medtronic and St Jude Medical. He serves as director of the Duke Neuro-Outcomes Center, which has received research funding from National Institutes of Health grant KM1 CA 156687, Medtronic, and St Jude Medical. Dr Hickey has consulted for Medtronic. He serves as Fellowship director of the Duke Movement Disorders Center, which has received funding from Medtronic. No other disclosures are reported.
Funding/Support: Grant KM1 CA 156687 from the National Institutes of Health.
Role of the Funder/Sponsor: The National Institutes of Health supplied funding for the collection, management, analysis, and interpretation of the data but had no role in the design and conduct of the study; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.