Introduction

The Chiari I malformation (CM-I) is the most common condition in the field of craniocervical junction [2, 4]. Surgery is the only treatment available for CM-I; however, multiple methods coexist for this procedure. Posterior fossa decompression with (PFDD) or without duraplasty (PFD) is the surgical treatment of choice [5, 18, 23]. With the trend towards minimally invasive techniques, the duraplasty portion of the procedure has become debatable for the elimination of it can lead to decreased complications. However, PFD is likely to cause higher rates of reoperation as it displays a tendency of inadequate depression [3]. Currently, the consensus on ideal surgical intervention for this condition has not been reached.

This paper will focus on the current literature examining outcomes following PFD compared to PFDD. We report a comprehensive systematic review and meta-analysis to evaluate the clinical advantages of the two methods for the treatment of CM-I.

Methodology

Search criteria

All full text randomized and non-randomized controlled trials, comparing the clinical outcomes of PFD and PFDD for the patients with CM-I in published studies, were included. Case reports of less than ten subjects, comments, letters, editorials, protocols, guidelines, animal studies, and cadaver articles were excluded.

Search strategy

The Medline, Embase, Cochrane library, Ovid, and CBM databases were searched for English-language articles published from May 1995 to May 2015. Unpublished studies were excluded. Prespecified search terms were “posterior fossa decompression”, “duraplasty,” and “chiari malformation type I.” Titles, abstracts, and subject headings were searched. The reference lists of all included articles and review papers were scrutinized for additional publications.

Search selection

Two reviewers independently assessed the titles and abstracts of each identified citation. The full text of potential articles were ordered and evaluated against the eligibility criteria. Any disagreements were resolved by discussion.

Data extraction

Each reviewer extracted data independently from each included paper. All data was tabulated onto a predefined spreadsheet. All articles were anonymized for author name, institution, journal title, and year of publication to blind reviewers during data extraction, appraisal, and analysis.

Outcome measures

The outcome measures were operation time, complications (neurological complications, cerebrospinal fluid (CSF) leak, wound infection, pseudomeningocele, and aseptic meningitis), clinical improvement, and recurrence rate.

Analysis

The mean differences and 95 % confidence intervals of each outcome were assessed by comparing PFDD and PFD groups, and the statistical heterogeneity was measured by using I 2 statistics. The I 2 test for heterogeneity was used to measure the proportion of total variation in study estimates due to heterogeneity rather than sampling error. If significant heterogeneity was found among studies based on interpretation of the I 2 test, a random effects model was applied. If no significant heterogeneity among studies was found, a fixed-effects model was applied. After this, the meta-analysis was carried out by using REVMAN software (version 5.0 for Windows. Copenhagen: The Nordic Cochrane Centre, the Cochrane Collaboration, 2008). p Values of less than 0.05 were considered statistically significant. We attempted to contact the original authors and inquire insufficient data of them. We also approached missing standard deviations for changes from the baseline by referring to Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (16.1.3.2).

Results

Search strategy

We found 864 potentially eligible articles of which 12 studies were included [7–10, 12, 13, 15, 16, 19–21, 25, 31] (Fig. 1). A total of 841 patients were included, 469 of which received PFD and the other 175 adopted PFDD. The follow-up periods were more than 6 months (Table 1).

Fig. 1
figure 1

A QUORUM chart

Full size image
Table 1 General condition of the 12 studies included in the meta-analysis
Full size table

Outcome measure

Mean operative time

There studies [15, 16, 21] recorded the mean operative time. A meta-analysis showed there is significant difference in the operative time [mean difference = −74.63, 95 % CI (−83.02, −66.25), p < 0.05] (Fig. 2).

Fig. 2
figure 2

A meta-analysis showed there is significant difference of the operative time [mean difference = −74.63, 95 % CI (−83.02, −66.25), p < 0.05]

Full size image

Complications

Twelve studies [7–10, 12, 13, 15, 16, 19–21, 25, 31] recorded the complications including neurological complications, CSF leak, wound infection, pseudomeningocele, and aseptic meningitis. There is significant difference in the complications between the two groups in favor of PFD [mean difference = 0.34, 95 % CI (0.19, 0.60), p < 0.05]. There is significant difference in the rate of CSF leak between the two groups [mean difference = 0.24, 95 % CI (0.07, 0.78), p < 0.05]. On the other hand, there is no significant difference between the two groups in the occurrence rate of neurological complications [mean difference = 0.25, 95 % CI (0.06, 1.10), p > 0.05], pseudomeningocele [mean difference = 0.25, 95 % CI (0.05, 1.21), p > 0.05], and aseptic meningitis [mean difference = 0.26, 95 % CI (0.05, 1.22), p > 0.05] and in the wound infection rate [mean difference = 0.86, 95 % CI (0.28, 2.68), p > 0.05] (Figs. 3 and 4).

Fig. 3
figure 3

There is significant difference in the complications between the two groups in favor of PFD [mean difference = 0.41, 95 % CI (0.30, 0.57), p < 0.05]. Specifically speaking, the occurrence rates of neurological complications [mean difference = 0.27, 95 % CI (0.07, 0.96), p < 0.05], CSF leak [mean difference = 0.16, 95 % CI (0.05, 0.51), p < 0.05], pseudomeningocele [mean difference = 0.10, 95 % CI (0.02, 0.53), p < 0.05], and aseptic meningitis [mean difference = 0.21, 95 % CI (0.05, 0.88), p < 0.05] are all significantly higher in the PFDD group compared to the PFD group. However, there is no significant difference in the wound infection rate between the two groups [mean difference = 0.86, 95 % CI (0.28, 2.68), p > 0.05]

Full size image
Fig. 4
figure 4

There is no difference in the clinical improvement rate between the two groups [mean difference = 0.55, 95 % CI (0.27, 1.10), p > 0.05]. Subgroup analysis shows that improvement rate is not significant both in children [mean difference = 0.49, 95 % CI (0.18, 1.35), p > 0.05] and adults [mean difference = 0.55, 95 % CI (0.23, 1.60), p > 0.05]

Full size image

Clinical improvement

Six studies [9, 10, 13, 16, 21, 31] recorded the clinical improvement rate. There is significant difference in the clinical improvement rate between the two groups in favor of the PFDD group [mean difference = 0.85, 95 % CI (0.73, 0.99), p < 0.05].

Critical appraisal

The population and study eligibility criteria are accurately defined, insuring comparable parameters are balanced and efficient in the study groups. Related research questions can be clearly addressed by appropriate selection and definition of the outcome measures. However, methodological limitations still exist in the literature (Table 2). Several factors may bias the findings of the current study. First, bias can be introduced in a retrospective review that does not have randomized, prospectively matched groups. Second, because both the practitioners and evaluators are surgeons, the design of the project cannot be absolutely blinded to which surgery was operated and clinical improvements such as numbness and tingling are hardly objective. Third, Some studies may reserve bony decompressions for mild cases or patients without syringomyelia and PFDD for the more severe cases which could incur bias. In both PFD and PFDD groups, the bony decompression range may also influence our results. Other confounding factors could be the operative decisions and techniques of the different surgeons. Some studies’ follow-up periods are quite short, and it may also affect the results. Multiple alternative factors, such as the implementation of tonsillar resection, the adoption of intraoperative ultrasound, and the different material used for duraplasty, could affect the outcome. This, therefore, permitted bias.

Table 2 Risk of bias assessment of included studies
Full size table

Discussion

The CM-I constitutes a group of congenital or acquired etiology that has descent of the cerebellar tonsils into the cervical spinal canal. The surgical indications of CM-I were usually headache (or tussive headache); neck, arm, or back pain; swallowing difficulties; drop attacks; or upper extremity numbness or tingling. The presence of a syrinx was also an indication for surgery when it occurred in the presence of the above symptoms [4, 5, 11, 20].

For these patients and neurosurgeons, the clinical improvement is the most priority for the surgical decision making. Generally, the postoperation symptomatic improvement has been reported in the range of 61.5–93 % from different studies [5, 6, 8, 16, 17]. Some studies indicated that PFDD is better at least for patients with syringomyelia and some report clinical results as comparable [21, 30]. According to our results, PFDD can achieve more satisfactory outcomes than PDF can in the clinical improvement between the two groups.

Some studies also reported the imaging improvement which is not as good as the clinical improvement [10, 12, 15, 31]. Though the compression has been relieved by surgical decompression, arachnoid scarring prevents the nervous structures to regain its normal position [10, 24]. However, the minimal enlargement of the subarachnoid spaces at the craniocervical junction resulting from bone decompression was sufficient to relieve the impact on the nervous structures and to improve CSF circulation. There is also no significant correlation between the reduction in syrinx size on MRI and the degree of clinical improvement [8, 10, 16, 29, 30]. Although clinical symptoms do not correlate to the presence or size of preoperative syringomyelia, there can be no doubt that a permanent postoperative reduction of syrinx size is an indicator of a sufficient decompression. On the contrary, the lower numbers for postoperative syrinx reductions in the bony decompression group must be seen as a prognostic indicator for worse long-term results. It implies that neurosurgeons should focus on both clinical outcome and radiological findings.

To better evaluate the effectiveness of PFD and PFDD in the management of CM-I, many researchers reported the re-operation rates [7, 9, 10, 15, 21, 22]. Once it occurs, the patients will suffer from tremendous pain and heavy economic burden. We hold the opinion that the re-operation may be caused by two main reasons: the persistence symptoms or the severe complications. Due to insufficient data, the different reasons of re-operation cannot be meta-analyzed in this paper. However, according to some research and our experience, more second surgeries had been done because of the recurrent syndromes such as nausea and vomiting rather than severe complications. Many studies observed a trend for lower recurrence rates in the PFDD group; whether this trend reaches the level of significance remains to be tested. For the patients who have recurrent clinical symptoms after PFD, a second PFDD can achieve better effect [1, 8, 15, 21, 26, 27].

For the other important aspect, PFD is associated with lesser rate of overall complications compared with PFDD especially in the CSF leak complication according to our results which reduces the re-operation to a great degree. The CSF-related complications include postoperative CSF leak, pseudomeningocele formation, meningitis, and scarring of the arachnoid leading to obstruction of CSF dynamics and foreign-body reaction [1, 3]. Most complications are related to CSF exposure to blood and muscle cellular debris as well as the use of dural graft. Thus, extradural decompression without duraplasty will be associated with lesser complication rates. Like previous studies [16, 23], we also found that the PFD group had shorter operative time than did the PFDD group. Besides, many researches have reported that PFD leads to shorter hospital stay lengths and less hospitalization costs [8, 17, 27].

CM-I is often associated with other medical conditions such as syringomyelia. Recent studies have reported an incidence rate as high as 70–80 % for syringomyelia [10, 28–30]. The range of the decrease rate of syringomyelia is wide (55–100 %) according to previous reports [8–10, 12, 16, 20, 26, 30]. Though not statistically significant, some findings also show higher rates of syrinx improvement in patients undergoing PFDD [23, 28–30]. Despite these findings, recently, a large sample study by Shweikeh [27] shows that CM-I patients with syringomyelia or hydromyelia were more often treated with PFD rather than PFDD. It is possible that the presence of syringomyelia does not currently influence treatment decisions.

Above all, for the criteria of the surgical decision making, we believe that for those with rapidly progressive symptoms or severe neurological deficits, previous research advises PFDD as the first option. While for patients with mild symptoms and no syringomyelia, PFD could be the first choice. If symptoms fail to improve, PFDD should be considered. According to our results, PFDD is better with improving rates but with more surgical trauma and higher rate of CSF leak. So, which is the preferable choice for those patients? As for the decompression surgery, our aim is to restore normal CSF dynamics at the level of the craniocervical junction. Thus, intraoperative ultrasonography may be a useful tool to aid the surgeon in deciding whether to adopt PFD or PFDD and even tonsillar shrinkage. Recently, more and more studies demonstrate that intraoperative ultrasonography can effectively guide the surgical decision making [14, 19, 32].

Conclusion

Although PFDD is related with longer operation time and higher CSF leak rate, it can still be considered as a preferable treatment option for most CM-I patients for its higher improvement rate. And PFDD tends to be worthy of consideration under the circumstance of failed PFD. More evidence from advanced multi-center studies are needed to provide illumination for the surgical decision making of CM-I.