Factors associated with implant failure in anterior cervical plating: retrospective analysis from a large cohort study

Abstract

Objectives

This study aims to retrospectively analyze the factors affecting implant failure in patients who underwent anterior cervical plating (ACP) following anterior cervical discectomy and fusion (ACDF).

Methods

Data from 332 patients who underwent ACDF and ACP at a single center between January 2011 and January 2023 were retrospectively reviewed. Patients who underwent ACP due to trauma, tumor, or infection, as well as those with missing data or incomplete follow-up, were excluded. Demographic and surgical data, such as age, sex, body mass index, preoperative and postoperative lordosis angles, type of screw used, screw length, bone quality, smoking and the number of fixed segments, were recorded. Implant failure (screw loosening, breakage, plate seperation) and pseudoarthrosis rate were evaluated. Logistic regression and Kaplan–Meier analysis were used for analyses.

Results

Implant failure was observed in 28 of the 332 patients (8.43%). The average follow-up duration was 36.8 months. Multivariate logistic regression analysis showed that greater fixation length (OR = 2.98, p < 0.0001), lower vertebral bone density measured by HU values (OR = 0.948, p < 0.0001), smoking (OR = 2.97, p = 0.032), and lower preoperative lordosis angle (OR = 0.873, p = 0.024) were independently associated with implant failure. Age, sex, type of screw, and postoperative lordosis angle had no significant effect on implant failure. According to Kaplan–Meier analysis, most failures occurred within the first six months. This study analyzes one of the largest anterior cervical plating series in the literature.

Conclusion

In anterior cervical plating, an increase in the number of fixed segments, reduced bone quality, smoking and a low preoperative lordosis angle are significant risk factors for implant failure. These factors should be considered in surgical planning.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00701-026-06928-x.

Introduction

Currently, the gold standard treatment for cervical disc disease is anterior cervical discectomy and fusion (ACDF) [6, 15]. To prevent complications such as graft failure, cage subsidence, cervical kyphosis, pseudoarthrosis, and graft dislocation following fusion, as well as to enhance fusion rates, anterior cervical plating (ACP) is recommended in addition to multilevel interbody fusion [3, 4, 13, 16].

Despite these advantages, implant-related complications remain a concern in patients undergoing ACP. System failure, including screw loosening, screw breakage, and plate fracture, has been reported in the literature and may negatively affect both clinical and radiological outcomes [9, 12, 14, 21, 22]. Several factors have been suggested to contribute to implant failure, including implant design, number of fused segments, and cervical sagittal alignment. mplant failure after ACP is likely multifactorial. Previous studies have suggested that construct length, bone quality, smoking status, cervical sagittal alignment, and implant-related factors may be associated with mechanical stability, fusion status, and implant-related complications [3, 8, 10, 12, 14, 17, 18, 21, 22]. However, these variables have often been evaluated in small series or heterogeneous cohorts, and their relative contribution remains unclear. Therefore, the variables evaluated in the present study were selected according to previous literature and clinical relevance.

Therefore, a clear gap persists in evaluating clinically relevant factors associated with implant failure in large, well-characterized patient populations.

The present study aims to address this gap by retrospectively analyzing one of the largest anterior cervical plating series reported to date. Using multivariate logistic regression analysis, this study seeks to identify factors independent risk factors associated with implant failure and to provide clinically applicable insights into their in vivo relevance.

Methods

This study is based on a retrospective analysis of patients who underwent anterior cervical plating (ACP) in addition to anterior cervical discectomy and fusion (ACDF) at our institution between January 2011 and January 2023. Approval was obtained from the local ethics committee for the use of anonymized patient data, and the study was conducted in accordance with the Declaration of Helsinki.

Patient data

The following data were collected for patients who underwent ACP surgery: age, sex, body mass index (BMI), smoking status, diabetes mellitus, preoperative cervical lordosis angle, postoperative cervical lordosis angle, bone quality, type and length of screws used in the system and number of plated segments. Patients were monitored for implant failure, including plate or screw breakage, screw loosening/pull-out and plate seperation. The duration of follow-up for each patient was also recorded.

Inclusion/exclusion criteria

Patients who underwent ACP fixation following ACDF during the specified period and had a minimum follow-up of 24 months were included in the study. All patients included in the study underwent ACDF for degenerative cervical spine disease such as cervical disc herniation, cervical spondylosis, or cervical myelopathy. Patients who underwent ACP due to trauma, tumor, or infection were excluded due to differences in underlying pathology. Additionally, patients who did not attend regular postoperative follow-ups, had missing or lost data, or had unavailable radiological records were excluded from the study.

Patients who did not attend regular postoperative follow-ups, had missing or lost data, or had unavailable radiological records were excluded from the study. Patients who could not be reached during the follow-up period despite scheduled clinical and radiological assessments were defined as lost to follow-up and were not included in the final analysis. Therefore, only patients with complete clinical and radiological follow-up data were included in the final cohort.

To evaluate potential selection bias, available baseline demographic, clinical, surgical, and radiological variables of patients included in the final cohort were compared with those of patients lost to follow-up.

Surgical technique

Routine anterior cervical microdiscectomy was performed. And PolyEtherAmide Ketone (PEEK)cages were filled with demineralized bone matrix (DBM) putty placed in discectomy site. After the discectomy and cage placement, the operated segments were stabilized using an anterior cervical plate. Monoaxial expandable or non-expandable screws were inserted monocortically into the vertebral bodies. During the study period, the surgeon utilized two different screw systems for fixation: self-locking expandable screws locked at the screw head (Fig. 1A) and non-expandable standard screws locked via the plate (Fig. 1B-C).

Fig. 1.

Implant and plating systems used. A Expandable screws: After the screw is fixed to the vertebral body, it expands to increase the bone-implant interface, aiming to enhance pull-out strength. B Standard non-expandable screw. C Locking system through the plate

Radiological Assesment: Pre- and post-operative lordosis angles for all patients were calculated from the lateral standing radiography. The degree of lordosis angles were calculated based on the posterior tangent method. The global cervical lordosis angle was defined as the angle between the two lines drawn parallel to the posterior surfaces of C2 and C7 (Fig. 2a). For the evaluation of operated segment mobility, the Cobb angle between the superior endplate of the upper instrumented vertebra (UIV) and the inferior endplate of the lower instrumented vertebra (LIV) on postoperative 6th month dynamic flexion–extension radiographs was considered (Fig. 2b). Motion greater than 2 degrees on radiographs obtained at the 6th postoperative month was evaluated as pseudarthrosis. Motion less than 2 degrees, together with the presence of bone trabeculation in the interbody area, was evaluated as fusion.

Fig. 2.

Radiological measurements used in the study. A Measurement of global cervical lordosis using the C2–C7 posterior tangential method. B Assessment of segmental mobility by measuring the Cobb angle between the superior endplate of the upper instrumented vertebra and the inferior endplate of the lower instrumented vertebra

In addition, plate fracture, screw fracture, screw loosening, screw pull-out, and plate separation detected on radiographs or CT scans were evaluated as implant failure.

To evaluate bone quality, Hounsfield Unit (HU) measurements were obtained from preoperative CT scans. HU values were measured at the C5 vertebral body using an elliptical region of interest placed in the trabecular bone on axial CT images while avoiding cortical bone and endplates. The average HU value was recorded for each patient.

Follow-up.

All operated patients underwent regular cervical x-ray follow-ups on the first postoperative day, at the 1 st, 6th, 12th, and 24th months, and these follow-ups were more frequent in patients with instrument failure.

The lordosis angle was measured preoperatively and postoperatively in all patients. The posterior tangent angle method was used to determine the lordosis angle.

The duration of follow-up was calculated from the date of surgery to the last available clinical or radiological assessment. Only patients with complete follow-up data were included in the final statistical analysis.

Statistic

Numerical data were be expressed as mean ± standard deviation (minimum–maximum) and categorical variables will be expressed as number (percentage %). The normal distribution of the data was be evaluated by Shapiro–Wilk test. Variables such as age, gender, smoking status, BMI, diabetes status, bone quality, preoperative lordosis, postoperative lordosis, type and length of screw used and number of segments fixed were compared between the group with failure and the control group. In these comparisons, Student's t-test or ANOVA was used if the data are normally distributed and Mann–Whitney U test or Kruskal–Wallis test was used if the data are not normally distributed. Univariate and multivariate logistic regression analysis were applied to identify the factors affecting the development of failure. Variables for the multivariable logistic regression model were selected a priori based on previous literature and clinical relevance. Considering the limited number of implant failure events, a partial model was used to reduce the risk of model overfitting. The selected variables included fixation length, C5 HU value, smoking status, screw type, screw length, and preoperative cervical lordosis. An additional exploratory sensitivity model including pseudarthrosis was also performed. Kaplan–Meier analysis was performed to examine the relationship between the time of development of failure.

Results

A total of 431 patients were initially identified. After applying the inclusion and exclusion criteria, 54 patients were excluded. Additionally, 45 patients were lost to follow-up. Therefore, a total of 332 patients with complete clinical and radiological data were included in the final analysis. In this final cohort, instrument failure occurred in 28 patients (8.43%). The mean follow-up period was 36.8 months (range 24–74) (Fig. 3). A baseline comparison between the included patients and the patients lost to follow-up was performed to assess possible selection bias. The available demographic, clinical, surgical, and radiological variables were compared between the groups and are presented in Supplementary Table 1. No major systematic difference was observed between the groups in the available baseline variables.

Fig. 3.

Various case examples of implant failure. A A patient who developed implant failure due to pull-out of screws at the C6 level after three-level fixation (C4-5–6) with ACP. B A patient with ACP at the C5-6–7 levels, where the plate separated from the vertebral body. C A patient who underwent ACP from C3-7, where the plate fractured at the C5-6 level

As a sensitivity estimate, if none of the 45 patients lost to follow-up had implant failure, the overall failure rate among eligible patients would be 7.4% (28/377). In an extreme worst-case scenario, if all patients lost to follow-up had implant failure, the failure rate would increase to 19.4% (73/377). Therefore, although baseline comparison did not show a major systematic difference between groups, the true implant failure rate may be affected by loss to follow-up.

The mean age of the patients at the time of surgery was 53.44 ± 10.69 years (range 26–76). Of the patients, 219 (65.96%) were female. When examining the number of segments treated with ACP fixation, 22 patients (6.62%) received 1 segments, 104 patients (31.32%) received 2 segments, 96 patients (28.91%) received 3 segments, 107 patients (32.22%) received 4 segments, and 3 patients (0.9%) received 5 segments of fixation. The preoperative cervical lordosis angle was measured as 11.22 ± 4.29 (range 2–27), while the postoperative cervical lordosis angle was measured as 19.83 ± 8.30 (range 1–45). The average lordosis gain was found to be 8.60 ± 7.63 (range −7 to 22). Self-locking expandable screws were used in 178 patients (53.61%), while plate-locking non-expandable screws were used in 154 patients (46.38%).

Comparison of the failure and control groups

The failure group consisted of 20 female and 8 male patients, while the control group contained 199 female and 105 male patients (p = 0.677). The average age in the failure group at the time of surgery was 53.57 ± 12.14 (range 32–72), while the control group at the time of surgery had an average age of 53.43 ± 10.57 (range 26–76) (p = 0.7871). Patients were also categorized according to age (< 60 years and ≥ 60 years). Implant failure occurred in 17 of 221 patients (7.7%) younger than 60 years and in 11 of 111 patients (9.9%) aged 60 years or older. This difference was not statistically significant (p = 0.49). Smoking was significantly more common in the failure group compared to the control group (35% vs 17%, p = 0.021). However, there was no significant difference between the groups regarding diabetes mellitus (25% vs 27%, p = 0.733). When patients were stratified according to the underlying pathology, implant failure occurred in 9 of 53 patients with cervical disc herniation (16.9%) and in 19 of 279 patients with cervical spondylotic stenosis (6.8%). This difference reached statistical significance (χ² = 4.72, p = 0.029). In the failure group, 1 patient received 1-segment fixation, 2 patients received 2-segment fixation, 9 patients received 3-segment fixation, 14 patients received 4-segment fixation, and 2 patients received 5-segment fixation. In the control group, 21 patients received 1-segment fixation, 102 patients received 2-segment fixation, 87 patients received 3-segment fixation, 93 patients received 4-segment fixation, and 1 patient received 5-segment fixation. A significant difference was found between the number of segments treated in the two groups (p < 0.0001). Longer fixations were more commonly observed in the failure group.

Regarding the type of screws and plates used, 19 patients in the failure group received self-locking expandable screws, while 9 patients received non-expandable via plate locked screws. In the control group, 159 patients received self-locking expandable screws, and 145 patients received non-expandable via plate locked screws. No significant difference was found between the failure and control groups in terms of the type of screw used (p = 0.114).

In the failure group, the preoperative lordosis angle was measured as 9.13 ± 5 (range 2–26), while in the control group it was measured as 11.42 ± 4.18 (p = 0.001). Postoperative lordosis in the failure group was 18.25 ± 9.05, while it was 19.98 ± 8.23 in the control group (p = 0.280). Hounsfield Unit (HU) measurements at the C5 vertebral body were significantly lower in the failure group compared to the control group (311 ± 49 vs 340 ± 58, p = 0.001).

Screw length did not differ significantly between the failure and control groups (14.5 ± 1.17 mm vs 14.97 ± 1.14 mm, p = 0.542) (Table 1).

Table 1.

Comprasion between failure and control group

	Failure Group (n = 28)	Control Group (n = 304)	p-value
Female	20 (71.42%)	199 (65.46%)	0.6771
Age	53.57 ± 12.14 (32–72)	53.43 ± 10.57 (26–76)	0.78712
BMI	28.20 ± 3.02 (23–35)	28.03 ± 2.89 (22–34)	0.8392
Smoking	10 (35%)	54 (17%)	0.0213
Diabetes	7 (25%)	85 (27%)	0.733
Underlying Pathology			0.0293
-Disc Herniation	9	44
-Spondylosis	19	260
Fixation Length
−1	1	21	< 0.00013
−2	2	102
−3	9	87
−4	14	93
−5	2	1
Screw Type
-Expandable	19	159	0.1143
-Non-Expandable	9	145
Pre-operative Lordosis	9.13 ± 5 (2–26)	11.42 ± 4.18 (3–24)	0.0012
Post-operative Lordosis	18.25 ± 9.05 (5–33)	19.98 ± 8.23 (1–38)	0.2802
Hounsfield Unit at C5 Vertebra	311 ± 49 (211–421)	340 ± 58 (194–402)	0.0012
Screw Length (milimeter)	14.5 ± 1.17 (12–16)	14.97 ± 1.14 (12–18)	0.542

Fisher Exact Test, 2 Mann Whitney-U, 3 Chi Square Test

Implant failure subtypes analysis

Implant failure events were further analyzed according to their subtype. Among the 28 failure cases, screw pull-out or plate loosening was the most common complication, observed in 15 patients (53.6%). Plate fracture occurred in 9 patients (32.1%), while screw breakage was observed in 4 patients (14.3%) (Table 2).

Table 2.

Subtypes of implant failure

Failure Type	Number of Cases	Percentage
Screw pull-out/Plate loosening	15	53%
Plate Fracture	9	32%
Screw Breakage	4	14%
Total	28	100%

Fusion status was also evaluated. Pseudarthrosis was detected in 21 of 28 patients (75%) in the failure group and in 28 of 304 patients (9.2%) in the control group. Pseudarthrosis was significantly more common in patients with implant failure (p < 0.0001). Revision surgery was performed in 19 of the 21 pseudarthrosis patients in the failure group, whereas 11 of the 28 pseudarthrosis patients in the control group underwent revision surgery.

Complication

No esophageal injuries were observed in this cohort. However, dysphagia developed in three patients who had plate separation.

Logistic reggression model for implant failure

Univariate logistic regression analysis was performed to evaluate the factors associated with implant failure (Table 3). Age, gender, BMI, diabetes mellitus, screw type, screw length, and postoperative lordosis were not significantly associated with implant failure. However, smoking (p = 0.025), lower Hounsfield Unit (HU) values of the C5 vertebral body (p < 0.0001), greater fixation length (p < 0.0001), and lower preoperative lordosis angle (p = 0.008) were significantly associated with implant failure.

Table 3.

Univariate logistic regression model for implant failure

Variable	B	S.E	Wald	df	Sig. (p-value)	(OR)Exp(B)	95% Cl
Age	−0.001	0.019	0.004	1	0.948	0.999	0.962–1.037
Gender (male)	−0.277	0.435	0.405	1	0.525	0.758	0.323–1.776
BMI	−0.009	0.068	0.16	1	0.272	0.991	0.867–1.132
Smoking	0.945	0.422	5.012	1	0.025	2.57	1.12–5.88
Diabetes	−0.152	0.455	0.112	1	0.738	0.859	0.352–2.09
Screw Type (expandable self locking)	−0.655	0.421	2.425	1	0.119	0.519	0.227–1.18
HU units of C5	−0.41	0.009	23.534	1	< 0.0001	0.959	0.942–0.976
Fixation Length	0.827	0.231	12.856	1	< 0.0001	2.288	1.46–3.59
Screw Length	−0.372	0.181	4.213	1	0.040	0.68	0.48–0.97
Pre-operative Lordosis	−0.143	0.054	7.063	1	0.008	0.866	0.78–0.96
Post-operative Lordosis	−0.027	0.028	0.032	1	0.291	0.973	0.92–1.03

Variables selected a priori based on previous literature and clinical relevance were included in the multivariable logistic regression model (Table 4). In the multivariate analysis, smoking (OR = 2.97, p = 0.032), lower HU values (OR = 0.948, p < 0.0001), greater fixation length (OR = 2.97, p < 0.0001), and lower preoperative lordosis angle (OR = 0.873, p = 0.024) remained associated with higher implant failure rate. Screw type and screw length showed a trend toward association but did not reach statistical significance in the multivariate model.

Table 4.

Multivariate logistic regression model for implant failure

Variable	B	S.E	Wald	df	Sig. (p-value)	OR (Exp(B))	95% Cl
Smoking	1.092	0.509	4.607	1	0.032	2.970	1.10–8.05
Screw Type (expandable self locking)	−0.941	0.514	3.348	1	0.067	0.390	0.14–1.07
HU units of C5	−0.52	0.010	30.386	1	< 0.0001	0.948	0.93–0.97
Fixation Length	1.160	0.314	13.687	1	< 0.0001	2.977	1.61–5.50
Screw Length	−0.456	0.239	3651	1	0.056	0.63	0.39–1.01
Pre-operative Lordosis	−0.136	0.060	5.065	1	0.024	0.873	0.78–0.98

Because pseudarthrosis may be associated with implant failure either as a contributing factor or as a parallel postoperative finding, an additional exploratory multivariate logistic regression analysis including pseudarthrosis was performed (Table 5). In this model, pseudarthrosis demonstrated a strong association with implant failure (OR = 32.7, p < 0.001). Greater fixation length (OR = 2.20, p = 0.028), lower C5 HU values (OR = 0.953, p < 0.001), and lower preoperative lordosis (OR = 0.845, p = 0.022) also remained significantly associated with implant failure. Smoking was no longer statistically significant after inclusion of pseudarthrosis in the model.

Table 5.

Exploratory multivariate logistic regression model including pseudarthrosis as an additional covariate

Variable	B	S.E	Wald	df	Sig. (p-value)	OR (Exp(B))	95% CI
Fixation length	0.789	0.360	4.800	1	0.028	2.201	1.087–4.459
Smoking	−0.747	0.621	1.447	1	0.229	0.474	0.140–1.600
HU units of C5	−0.048	0.010	24.456	1	< 0.001	0.953	0.935–0.971
Preoperative lordosis	−0.168	0.073	5.271	1	0.022	0.845	0.732–0.976
Pseudarthrosis	3.488	0.598	34.024	1	< 0.001	32.723	10.135–105.649

Kaplan meier analysis

According to the Kaplan–Meier analysis, implant failure occurred in a total of 4 cases in the first month, with the proportion of patients without implant failure found to be 98.8% (Std. Error: 0.006). In the third month, failure was observed in 13 cases, reducing the rate to 95.5% (Std. Error: 0.011). By the sixth month, the number of failure cases increased to 21, with a survival rate of 93.7% (Std. Error: 0.013). At the end of the first year, implant failure was detected in 26 cases, with the rate determined as 92.2% (Std. Error: 0.015). By the end of the second year, the number of failure cases reached 27, with a rate of 91.9% (Std. Error: 0.015). At the end of the fifth year, failure was observed in 28 cases, and the rate was found to be 91.6% (Std. Error: 0.015) (Table 6, Fig. 4).

Table 6.

Survival table with kaplan–meier analysis

Time	Cumulative Proportion Surviving at the Time		N of Cumulative Events
	Estimate	Std.Error
First Month	0.988	0.006	4
Third Month	0.955	0.011	13
Sixth Month	0.937	0.013	21
First Year	0.922	0.015	26
Second Year	0.919	0.015	27
Five Year	0.916	0.015	28

Fig. 4.

Kaplan–Meier survival curve, demonstrating the log survival function of implants over the study period

Discussion

In this study, the factors influencing implant failure in patients who underwent anterior cervical plating (ACP) following anterior cervical discectomy and fusion (ACDF) were retrospectively analyzed. Logistic regression analysis revealed that age, sex, BMI, diabetes mellitus status, screw type, screw length and postoperative lordosis angle had no significant effect on implant failure. However, greater fixation length was independently associated with implant failure. Additionally, lower preoperative lordosis angle were associated with higher implant failure rates. Poor bone quality and smoking are also other two risk factor for implant failure. Kaplan–Meier analysis showed that implant failure predominantly occurred within the first six months, while implant integrity was preserved at a rate of 91.6% by the fifth year. These findings suggest that fixation length, preoperative alignment, smoking, and bone quality may be clinically relevant with higher implant failure.

Possible biomechanical interpretations: (1) One possible explanation for the association between greater fixation length and implant failure is the altered biomechanical environment observed in longer constructs. Longer fixation may increase construct rigidity and cyclic loading on the implant system. However, fixation length may also reflect underlying disease severity, multilevel degeneration, or more complex surgical indications rather than a purely independent biomechanical factor [1, 11, 20, 22]. (2) Similarly, lower preoperative lordosis may be associated with implant failure through several mechanisms. Patients with kyphotic or low-lordosis alignment may require greater surgical correction and may represent more advanced degenerative disease. Therefore, preoperative alignment may reflect both biomechanical factors and underlying disease complexity. (3) The early postoperative period, particularly the first six months, may represent a vulnerable phase for implant stability because solid osseointegration and fusion have not yet fully developed [2, 5, 7]. During this period, repetitive loading may increase the risk of screw loosening or plate-related complications. This interpretation is consistent with our Kaplan–Meier analysis, which demonstrated that most implant failures occurred within the first six months after surgery. (4) Bone quality is an important determinant of screw fixation strength. Lower vertebral bone density may reduce screw purchase and increase the risk of screw loosening or pull-out [10]. In our study, the failure group demonstrated significantly lower HU values compared to the control group. This finding suggests that reduced bone quality may contribute to implant failure in anterior cervical plating. (5) Different implant failure subtypes may involve different biomechanical mechanisms; however, the present study was not specifically designed to evaluate these mechanisms separately. (6) Fusion status was also evaluated. Pseudarthrosis was detected in 21 of 28 patients (75%) in the failure group and in 28 of 304 patients (9.2%) in the control group. Pseudarthrosis was significantly more common in patients with implant failure (p < 0.0001).

Pseudarthrosis was significantly more common in patients with implant failure. Additional exploratory multivariate analysis including pseudarthrosis also demonstrated a strong association between non-union and implant failure. However, due to the retrospective design of the study, the temporal and causal relationship between pseudarthrosis and implant failure could not be definitively established. Implant failure may occur secondary to persistent motion caused by non-union, whereas in some patients implant instability itself may contribute to impaired fusion. Interestingly, smoking lost statistical significance after inclusion of pseudarthrosis in the model, suggesting that part of the smoking-related effect may be mediated through impaired fusion biology. Therefore, these findings should be interpreted as a strong association rather than proof of a direct causal sequence.

Clinical Implication: From a clinical perspective, the identified factors may help surgeons recognize patients at higher risk for implant failure. Patients requiring long-segment fixation, those with reduced bone quality, active smokers, and patients with low preoperative lordosis may benefit from closer postoperative follow-up, especially during the first six months after surgery. In these patients, careful construct planning, optimization of bone health, smoking cessation strategies, and close attention to fusion status may help reduce implant-related complications. In long constructs or patients with poor bone quality, additional preventive strategies such as stronger fixation planning, careful screw trajectory and length selection, and more frequent radiological follow-up may be considered.

Expandable screws theoretically have higher pull-out strength due to the increased bone-implant surface area [17, 18]. However, in our study, no significant difference was observed between expandable screws and regular screws in terms of implant failure.

In a series of 135 patients published by Lowery et al., the implant failure rate was reported to be 35% [12]. This study demonstrated that different plate types had varying effects on failure. It was reported that the segment length had either positive or negative effects on failure, depending on the type of plate used. However, in our study, regardless of the implant type, an increase in the number of segments was found to increase the risk of failure. In another study by Stulik et al., in a series of 132 patients, implant failure was reported in only 4 patients (3%) [21]. In the finite element model conducted by Wang et al., it was demonstrated that including the C6-7 and C7-T1 levels in anterior plating would weaken the system biomechanically, as shown in their model [23]. Song et al. reported that plate augmentation in ACDF constructs may improve biomechanical stability and contribute to higher fusion rates compared with cage-only techniques [19].

From a practical point of view, several strategies may help reduce implant failure. In patients who require long-segment fixation, careful implant selection and appropriate construct design are important to reduce excessive mechanical stress on the plate system. Bone quality should also be considered, especially in patients with osteoporosis, since poor bone density may reduce screw pull-out strength. In addition, achieving solid fusion is important because persistent motion due to pseudarthrosis may increase repeated loading on the hardware. Therefore, careful surgical technique, appropriate construct length, and attention to bone quality may help reduce implant-related complications.

This study has several limitations. First, it has the inherent limitations of a retrospective design. Because the study was not randomized, selection bias could not be fully excluded, and retrospective data collection may have introduced documentation bias. DEXA measurements were not available for all patients; therefore, C5 Hounsfield Unit values were used as a surrogate marker of bone quality. In addition, the severity of degenerative changes was not systematically graded in this dataset. The study was conducted at a single center, and the results may have been influenced by surgeon experience and specific surgical techniques, which may limit generalizability. Another limitation is the heterogeneity of surgical indications, as both cervical disc herniation and cervical spondylotic stenosis were included. Patients lost to follow-up were excluded from the final analysis. Although baseline comparison did not demonstrate a major systematic difference between the included patients and those lost to follow-up, implant failure status could not be confirmed in these patients. Therefore, residual selection bias and limited external validity cannot be completely excluded. In addition, some identified factors, such as fixation length and preoperative alignment, may partly reflect disease severity or surgical indication rather than direct causal biomechanical mechanisms. Therefore, these findings should be interpreted as associations rather than definitive proof of causality. Finally, although the overall cohort was relatively large, only 28 implant failure events were observed. Therefore, the multivariable regression models may be limited by statistical power and potential overfitting, especially when several covariates are included. For this reason, the regression results should be interpreted as association analyses rather than as a formal predictive model. However, the strengths of this study include the long follow-up period and the high number of patients. These factors enhance the reliability of the statistical results and ensure that the study makes a significant contribution to the literature.

Conclusion

This study highlights that greater fixation length and lower preoperative lordosis were associated with implant failure in anterior cervical plating. Smoking and poor bone quality were also associated with higher implant failure rates. The findings emphasize the importance of carefully considering fixation length, bone quality and preoperative alignment when planning anterior cervical discectomy and fusion with plating. Despite these factors, implant integrity remained high beyond the initial six months, supporting the long-term effectiveness of ACP in most patients. These findings should be interpreted cautiously because some variables may also reflect underlying disease severity and surgical complexity.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

All authors contributed to the study conception and design. Material preparation, data collection were performed by Ahmet Serhat Eroğlu. Data analysis were performed by Oğuz Kağan Demirtaş. The first draft of the manuscript was written by Oğuz Kağan Demirtaş and all authors commented on previous versions of the manuscript. All authors read, contribute and approved the final manuscript.

Funding

No-funding was received for this research.

Data availability

Data not available due to ethical and legal restrictions.

Declarations

Ethical approval

Approval was obtained from the Medicalpark Özel Ordu Hospital IRB for the use of anonymized patient data, and the study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Informed consent was obtained from all individual participants included in the study.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Competing interests

The authors declare no competing interests.