Anterior approaches to the lumbar spine have been used for more than a century, beginning with a transperitoneal approach by Muller and colleagues in 1906.1,2 Iwahara was the first to publish on a retroperitoneal approach in 1944, followed by Lane and Moore in 1948 with their publication of an early case series of patients treated with anterior lumbar interbody fusion (ALIF) for isolated lumbar disc disease.1,3,4 In 1984, the first lumbar total disc replacement (TDR) was performed as a motion-preserving alternative to ALIF.5 The theoretical benefits of an anterior approach over other approaches to the lumbar spine include the avoidance of neural elements, less blood loss, and the unique biomechanical advantage provided by its ability to restore intervertebral height, sagittal balance, and lumbar lordosis by virtue of the relatively large cage inserted.1
The anterior approach increased in popularity in 2002 after the FDA’s approval of recombinant human bone morphogenetic protein–2 (rhBMP-2) for use in L4/5 and L5/S1 ALIF with Medtronic’s LT cage, secondary to the reportedly superior fusion rates.1 By stimulating pluripotent stem cells in the bleeding cancellous bone, rhBMP-2 induces osteogenesis and promotes solid fusion.1,6 Whereas original industry-sponsored research on rhBMP-2 reported 0% rhBMP-2–associated adverse effects,7 subsequent research on its safety demonstrated its use to be associated with higher rates of complications than the standard iliac crest bone autograft (ICBG), leading the FDA to issue a warning in 2008 on the use of rhBMP-2 in both an on-label and off-label manner.8 Specifically regarding use in ALIF procedures, this warning addressed the risk of osteolysis, retrograde ejaculation (RE), and other urological complications, as well as delayed infection.8 The use of rhBMP-2 has continued, with surgeons now adopting precautions in its handling and placement, for instance the utilization of a surgical barrier to ensure confinement within the cage.9
With exposure of the great vessels as a major step, complications of an anterior approach have the potential to be devastating. However, the overall rate of complication has been kept relatively low, with vascular injury occurring in less than 15% of cases.1,10 RE is another feared complication of the anterior approach that is inconsistently reported in the literature, leading to difficulty counseling male patients on the inherent risks of the surgical approach.1,9–11
The pathophysiology of RE in anterior lumbar surgery is secondary to damage to the hypogastric plexus traversing the prevertebral space near the L5/S1 interspace. The resulting disruption of sympathetic function causes relaxation of the internal bladder sphincter at time of ejaculation, resulting in retrograde flow of ejaculate into the bladder.1,9–11 The risk of RE following anterior lumbar surgery has received special attention, with much debate in the literature over its incidence due to the subjectivity in self-reporting and the lack of published studies incorporating objective data such as postejaculation urinalysis.12 Likewise, risk factors have been difficult to delineate.
RE is a strictly male complication, but most studies of complications in anterior lumbar surgery are not solely investigating RE, so there is a paucity of literature reporting data on surgical factors segregated by sex. The uncertainty surrounding this risk and its implications on reproductive potential have led some surgeons to renounce ALIF in men with child-siring potential. The present systematic review and pooled analysis is intended to assess the incidence and risk factors for development of RE following first-time open modern anterior lumbar surgery.
Methods
Search Strategy
A systematic MEDLINE search via PubMed was performed for published articles related to RE following first-time open anterior lumbar surgery in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (Fig. 1).13 The search was limited to peer-reviewed clinical studies published in the English language between the inception of PubMed and April 2020, in which the full text was available. The search criteria were subsequently defined as follows:
Anterior lumbar [All Fields] AND interbody fusion [All Fields] AND retrograde ejaculation [All Fields]
Anterior lumbar [All Fields] AND arthroplasty [All Fields] AND retrograde ejaculation [All Fields]
Anterior lumbar [All Fields] AND retrograde ejaculation [All Fields]
Anterior lumbosacral [All Fields] AND retrograde ejaculation [All Fields]
Flowchart depicting the study selection process.
The searches yielded 44, 9, 63, and 14 publications, respectively. Titles and abstracts were reviewed. Editorials, abstracts, conference presentations, literature reviews, and meta-analyses were excluded. Of these 130 search results, 31 publications were selected for full-text review. Studies on laparoscopic techniques and studies not reporting quantitative RE data were excluded. If multiple studies reported on the same cohort of patients, inclusion was limited to the most complete publication with the most recent data, although details may have been cross-referenced with additional papers on a given cohort published by the same author group, and this is cited where appropriate. Minimum follow-up time was not used as one of the exclusionary criteria, given the nature of RE as an early complication. Eighteen papers met inclusion criteria. Levels of evidence ratings were assigned to each paper based on criteria described by Wright et al.14
Data Extraction
Data were extracted by 1 author. All articles were reviewed for extraction of the following data for the men included in the study: years during which surgeries were performed, number of men included, retroperitoneal versus transperitoneal approach, single-level versus multilevel procedure, procedure performed (ALIF vs TDR), use of rhBMP-2, number of men who reported RE postoperatively, and number of men who reported RE persistence at final follow-up.
Statistical Analysis
A pooled analysis of the individual studies was performed rather than a true meta-analysis, given the rarity of the event and the small sample size of individual studies. All studies meeting the inclusion criteria were included in the analysis of incidence of RE. Studies reporting outcome at final follow-up in patients with RE and studies in which no events of RE occurred were included in the analysis of incidence of persistent RE and rate of resolution of RE. Because many studies did not report on surgical details segregated by sex, our analysis of risk factors included only the studies that strictly included men or clarified the number of men with a given risk factor. A 2-sided Fisher’s exact test or chi-square test was used to test for statistically significant differences between comparison groups. A p value ≤ 0.05 was considered statistically significant. All statistical analysis was performed with SPSS version 26.0 (IBM Corp.).
Results
Study Characteristics
Eighteen publications met inclusion criteria and were used for data extraction. Four of these provided level I evidence, 2 provided level III evidence, and 12 provided level IV evidence. Characteristics of the included studies, totaling 2503 male patients, are summarized in Table 1. Six of the 18 included studies did not provide details on how RE was detected.15,16,19,20,25,28 The average RE rate for these studies was 1.0% (range 0%–3.8%). Only 1 study used laboratory evaluation in all patients, detecting a 9.8% objective rate of RE at 3 months and a 4.9% objective rate at 6 months.12 Four studies report using questionnaires, with an average RE rate of 5.6% (range 1.5%–9.8%).9,11,12,33 Four studies report directly querying patients on RE symptoms (decreased volume/absence of ejaculate, infertility), with an average RE rate of 2.8% (range 0%–8.4%).11,18,21,26 These studies did not specify whether further workup was pursued. Furthermore, it was not specified whether isolated infertility was counted as RE. Three studies with an average RE rate of 2.9% (range 0%–6.4%) note that they did not specifically ask about RE but discussed the complication as part of the informed consent and then questioned patients postoperatively about any complications.17,23,29 Other methods noted included access surgeons retrospectively querying spine surgeons about their patients’ complications (RE rate 0.3%)27 and questioning patients about adverse events, although it was not specified what this questioning entailed (RE rate 4.1%).22
Study characteristics and reported incidence of RE
| Authors & Year | Methods Reported | Average Age in Yrs (range) | No. of Pts | No. of Men | Cases of RE | RE Rate | Persistent Cases | Rate of Resolution |
|---|---|---|---|---|---|---|---|---|
| Amaral et al., 201715 | Methods not indicated | 44 | 87 | 37 | 0 | 0.0% | 0 | NA |
| Arlet et al., 200616 | Methods not indicated | 39.9 (26–57) | 21 | 5 | 0 | 0.0% | 0 | NA |
| Brau, 200227 | Access surgeons queried spine surgeons on their pts | (19–84) | 686 | 345 | 1 | 0.3% | 1 | 0.0% |
| Burkus et al., 200517 | Not directly questioned | Study group: 40.2; control group: 43.6 | 131 | 51 | 0 | 0.0% | 0 | NA |
| Comer et al., 201211 | Questionnaire & direct questioning | (19–65) | 472 | 472 | 17 | 3.6% | 10 | 41.2% |
| Edgard-Rosa et al., 201218 | Direct questioning | 43 (20–73) | 469 | 337 | 0 | 0.0% | 0 | NA |
| Fischer et al., 201433 | Questionnaire | 49.2 | 58 | 31 | 1 | 3.2% | NR | NR |
| Gornet et al., 201123 | Not directly questioned | (18–70) | 577 | 291 | 6 | 2.1% | 2 | 66.7% |
| Jarrett et al., 200929 | Not directly questioned | 50 (16–85) | 265 | 110 | 7 | 6.4% | NR | NR |
| Kang et al., 200919 | Methods not indicated | 56 (34–79) | 412 | 114 | 0 | 0.0% | 0 | NA |
| Lindley et al., 201221 | Direct questioning | (23–76) | 95 | 95 | 8 | 8.4% | 6 | 25.0% |
| Loguidice et al., 198825 | Methods not indicated | 41 (21–63) | 85 | 44 | 1 | 2.3% | NR | NR |
| Malham et al., 20149 | Questionnaire | 45.3 (23–70) | 131 | 67 | 1 | 1.5% | 1 | 0.0% |
| Mobbs et al., 201628 | Methods not indicated | 57.6 (25–86) | 227 | 105 | 4 | 3.8% | 1 | 75.0% |
| Newman & Grinstead, 199226 | Direct questioning | (26–58) | 36 | 15 | 1 | 6.7% | 0 | 100.0% |
| Quraishi et al., 201320 | Methods not indicated | 43 (10–73) | 304 | 197 | 0 | 0.0% | 0 | NA |
| Sasso et al., 200322 | Pts questioned about adverse events | NR | 146 | 146 | 6 | 4.1% | 3 | 50.0% |
| Tepper et al., 201312 | Lab evaluation & questionnaire | (24–56) | 41 | 41 | 4† | 9.8% | 2 | 50.0% |
| Total | 4243 | 2503 | 57 | 2.3% | 26 | 45.8%* |
NA = not applicable; NR = not reported; Pts = patients.
Of the 48 reported cases of RE for which condition at final follow-up was documented, 22/48 had resolved (45.8%).
Four of 41 patients had RE as detected by laboratory evaluation, whereas 15/41 reported RE on questionnaire.
Eighteen publications met inclusion criteria and were used for data extraction. Four of these provided level I evidence, 2 provided level III evidence, and 12 provided level IV evidence. Characteristics of the included studies, totaling 2503 male patients, are summarized in Table 1. Six of the 18 included studies did not provide details on how RE was detected.15,16,19,20,25,28 The average RE rate for these studies was 1.0% (range 0%–3.8%). Only 1 study used laboratory evaluation in all patients, detecting a 9.8% objective rate of RE at 3 months and a 4.9% objective rate at 6 months.12 Four studies report using questionnaires, with an average RE rate of 5.6% (range 1.5%–9.8%).9,11,12,33 Four studies report directly querying patients on RE symptoms (decreased volume/absence of ejaculate, infertility), with an average RE rate of 2.8% (range 0%–8.4%).11,18,21,26 These studies did not specify whether further workup was pursued. Furthermore, it was not specified whether isolated infertility was counted as RE. Three studies with an average RE rate of 2.9% (range 0%–6.4%) note that they did not specifically ask about RE but discussed the complication as part of the informed consent and then questioned patients postoperatively about any complications.17,23,29 Other methods noted included access surgeons retrospectively querying spine surgeons about their patients’ complications (RE rate 0.3%)27 and questioning patients about adverse events, although it was not specified what this questioning entailed (RE rate 4.1%).22
Because many of the included studies did not focus on RE and thus the methods surrounding RE detection were often unclear, bias was likewise difficult to assess on an individual study basis. Given the retrospective nature of many of the studies, we believe information bias to be prevalent. Additionally, these studies were probably susceptible to selection bias throughout time as some surgeons began to perform fewer ALIFs on men with child-siring potential.
Incidence of RE
The incidence of RE varied among the studies from 0.0%15–20 to 9.8%12 (Table 1). However, it must be noted that those 9.8% were diagnosed based on laboratory criteria, whereas 41.7% of patients who filled out the questionnaire self-reported a decrease in ejaculate volume.12 For pooled analysis, the present study used the number diagnosed objectively because our intention was to approximate the true incidence as closely as possible. In total, the studies reported 57 events of RE in 2503 men (2.3%). This resulted in 26 cases of persistent RE at final follow-up for the 2318 men included in studies that reported duration of symptoms or had no cases of RE, for a pooled incidence of persistent RE of 1.1% and a rate of resolution of 45.8%. Final follow-up ranged from 6 to 62 months, although the average could not be determined (Table 2).
Reported information on time of resolution or persistence of RE symptoms, ordered by rate of resolution
| Authors & Year | Cases of RE | No. Resolved | Time of Resolution | No. Persistent | Interval of Persistence | Rate of Resolution |
|---|---|---|---|---|---|---|
| Brau, 200227 | 1 | 0 | NA | 1 | 6 mos | 0.0% |
| Malham et al., 20149 | 1 | 0 | NA | 1 | 12 mos | 0.0% |
| Lindley et al., 201221 | 8 | 2 | NR | 6 | NR | 25.0% |
| Comer et al., 201211 | 17 | 7 | 7 resolved by 12 mos | 10 | 17/17 persisted at 6 mos; 10/17 persisted at 12 mos | 41.2% |
| Sasso et al., 200322* | 6 | 3 | 2 resolved by 12 mos (1 rhBMP-2, 1 ctrl); 1 resolved by 48 mos (rhBMP-2) | 3 | 3/6 persisted at final FU, minimum 2 yrs (3/5 rhBMP-2 vs 0/1 ctrl) | 50.0% |
| Tepper et al., 201312 | 4 | 2 | 1 each resolved as evidenced by pregnancy at 2 yrs & 3 yrs, respectively | 2 | NR† | 50.0% |
| Gornet et al., 201123* | 6 | 4 | 2 resolved by 6 mos (2 rhBMP-2); 1 resolved by 12 mos (ctrl); 1 resolved by 24 mos (ctrl) | 2 | 2/6 persisted at 24 mos (0/2 rhBMP-2 vs 2/4 ctrl) | 66.7% |
| Mobbs et al., 201628 | 4 | 3 | 3 resolved by 3 mos | 1 | NR | 75.0% |
| Newman & Grinstead, 199226 | 1 | 1 | Resolved by 4 mos | 0 | NA | 100.0% |
Ctrl = control; FU = follow-up.
Data for this cohort were cross-referenced between this publication and that of Burkus et al., 2013.24
All cases diagnosed by laboratory criteria at 3–6 months; no further information given on follow-up.
Risk Factors
The incidences of RE associated with various surgical risk factors are shown in Tables 3–7. Four of the included studies were composed strictly of men and thus could be included in all risk factor analyses.11,12,21,22 Another 2 studies17,23 included women but were outlined in a meta-analysis by the same group of authors that was excluded from the present study but was able to be cross-referenced where necessary.24 Other studies were included as appropriate in situations wherein the authors explicitly noted the number of men with a given risk factor.
Cases of RE by retroperitoneal versus transperitoneal approach
| Authors & Year | RP Group | TP Group | Statistical Significance | ||||
|---|---|---|---|---|---|---|---|
| No. | RE Events | RE Rate | No. | RE Events | RE Rate | ||
| Burkus et al., 200517*† | 42 | 0 | 0.0% | 4 | 0 | 0.0% | No |
| Comer et al., 201211 | 472 | 17 | 3.6% | 0 | NA | NA | NA |
| Gornet et al., 201123* | 267 | 5 | 1.9% | 24 | 1 | 4.2% | No |
| Kang et al., 200919 | 114 | 0 | 0.0% | 0 | NA | NA | NA |
| Lindley et al., 201221 | 95 | 8 | 8.4% | 0 | NA | NA | NA |
| Loguidice et al., 198825 | 44 | 1 | 2.3% | 0 | NA | NA | NA |
| Newman & Grinstead, 199226 | 15 | 1 | 6.7% | 0 | NA | NA | NA |
| Sasso et al., 200322 | 116 | 2 | 1.7% | 30 | 4 | 13.3% | Yes |
| Tepper et al., 201312 | 41 | 4 | 9.8% | 0 | NA | NA | NA |
| Total | 1206 | 38 | 3.2% | 58 | 5 | 8.6% | Yes (p = 0.043) |
RP = retroperitoneal; TP = transperitoneal.
Data for this cohort were cross-referenced between this publication and that of Burkus et al., 2013.24
Five patients could not be classified.
Cases of RE by procedure performed
| Authors & Year | ALIF Group | TDR Group | Statistical Significance | ||||
|---|---|---|---|---|---|---|---|
| No. | RE Events | RE Rate | No. | RE Events | RE Rate | ||
| Amaral et al., 201715 | 37 | 0 | 0.0% | 0 | NA | NA | NA |
| Arlet et al., 200616 | 5 | 0 | 0.0% | 0 | NA | NA | NA |
| Brau, 200227 | 345 | 1 | 0.3% | 0 | NA | NA | NA |
| Burkus et al., 200517 | 51 | 0 | 0.0% | 0 | NA | NA | NA |
| Comer et al., 201211 | 472 | 17 | 3.6% | 0 | NA | NA | NA |
| Gornet et al., 201123 | 86 | 2 | 2.3% | 205 | 4 | 2.0% | No |
| Lindley et al., 201221 | 54 | 4 | 7.4% | 41 | 4 | 9.8% | No |
| Loguidice et al., 198825 | 44 | 1 | 2.3% | 0 | NA | NA | NA |
| Malham et al., 20149 | 67 | 1 | 1.5% | 0 | NA | NA | NA |
| Mobbs et al., 201628 | 105 | 4 | 3.8% | 0 | NA | NA | NA |
| Newman & Grinstead, 199226 | 15 | 1 | 6.7% | 0 | NA | NA | NA |
| Sasso et al., 200322 | 146 | 6 | 4.1% | 0 | NA | NA | NA |
| Tepper et al., 201312 | 41 | 4 | 9.8% | 0 | NA | NA | NA |
| Total | 1468 | 41 | 2.8% | 246 | 8 | 3.3% | No (p = 0.84) |
Cases of RE by single- versus multilevel exposure
| Authors & Year | Single-Level Group | Multilevel Group | Statistical Significance | ||||
|---|---|---|---|---|---|---|---|
| No. | RE Events | RE Rate | No. | RE Events | RE Rate | ||
| Burkus et al., 200517* | 51 | 0 | 0.0% | 0 | NA | NA | NA |
| Comer et al., 201211 | 302 | 10 | 3.3% | 170 | 7 | 4.1% | No |
| Gornet et al., 201123* | 291 | 6 | 2.1% | 0 | NA | NA | NA |
| Jarrett et al., 200929 | 67 | 5 | 7.5% | 43 | 2 | 4.7% | No |
| Lindley et al., 201221 | 41 | 5 | 12.2% | 54 | 3 | 5.6% | No |
| Sasso et al., 200322 | 146 | 6 | 4.1% | 0 | NA | NA | NA |
| Tepper et al., 201312 | 26 | 2 | 7.7% | 15 | 2 | 13.3% | No |
| Total | 924 | 34 | 3.7% | 282 | 14 | 5.0% | No (p = 0.38) |
Data for this cohort were cross-referenced between this publication and that of Burkus et al., 2013.24
Cases of RE by rhBMP-2 use
| Authors & Year | rhBMP-2 Group | Ctrl | Statistical Significance | ||||
|---|---|---|---|---|---|---|---|
| No. | RE Events | RE Rate | No. | RE Events | RE Rate | ||
| Amaral et al., 201715 | 0 | NA | NA | 37 | 0 | 0.0% | NA |
| Arlet et al., 200616 | 0 | NA | NA | 5 | 0 | 0.0% | NA |
| Brau, 200227 | 0 | NA | NA | 345 | 1 | 0.3% | NA |
| Burkus et al., 200517* | 32 | 0 | 0.0% | 19 | 0 | 0.0% | No |
| Comer et al., 201211 | 239 | 15 | 6.3% | 233 | 2 | 0.9% | Yes |
| Gornet et al., 201123* | 86 | 2 | 2.3% | 205 | 4 | 2.0% | No |
| Lindley et al., 201221 | 54 | 4 | 7.4% | 41 | 4 | 9.8% | No |
| Loguidice et al., 198825 | 0 | NA | NA | 44 | 1 | 2.3% | NA |
| Malham et al., 20149 | 67 | 1 | 1.5% | 0 | NA | NA | NA |
| Mobbs et al., 201628 | 0 | NA | NA | 105 | 4 | 3.8% | NA |
| Newman & Grinstead, 199226 | 0 | NA | NA | 15 | 1 | 6.7% | NA |
| Sasso et al., 200322 | 78 | 5 | 6.4% | 68 | 1 | 1.5% | No |
| Tepper et al., 201312 | 21 | 2 | 9.5% | 20 | 2 | 10.0% | No |
| Total | 577 | 29 | 5.0% | 1137 | 20 | 1.8% | Yes (p < 0.001) |
Data for this cohort were cross-referenced between this publication and that of Burkus et al., 2013.24
Cases of RE by rhBMP-2 use (time segregated)
| Authors & Year | Study Period | rhBMP-2 Group | Ctrl | Statistical Significance | ||||
|---|---|---|---|---|---|---|---|---|
| No. | RE Events | RE Rate | No. | RE Events | RE Rate | |||
| Arlet et al., 200616 | NR | 0 | NA | NA | 5 | 0 | 0.0% | NA |
| Brau, 200227 | 1997–2000 | 0 | NA | NA | 345 | 1 | 0.3% | NA |
| Burkus et al., 200517* | 2000–2001 | 32 | 0 | 0.0% | 19 | 0 | 0.0% | No |
| Comer et al., 201211† | 2002–2008 | 179 | 13 | 7.3% | 174 | 1 | 0.6% | Yes |
| Gornet et al., 201123 | 2003–2004 | 86 | 2 | 2.3% | 205 | 4 | 2.0% | No |
| Loguidice et al., 198825 | 1981–1985 | 0 | NA | NA | 44 | 1 | 2.3% | NA |
| Newman & Grinstead, 199226 | NR | 0 | NA | NA | 15 | 1 | 6.7% | NA |
| Sasso et al., 200322 | 1998–1999 | 78 | 5 | 6.4% | 68 | 1 | 1.5% | No |
| Total (pre-2008) | 375 | 20 | 5.3% | 875 | 9 | Yes (p < 0.001) | ||
| Amaral et al., 201715 | 2009–2016 | 0 | NA | NA | 37 | 0 | 0.0% | NA |
| Comer et al., 201211† | 2008–2011 | 60 | 2 | 3.3% | 59 | 1 | 1.7% | No |
| Malham et al., 20149 | 2009–2012 | 67 | 1 | 1.5% | 0 | NA | NA | NA |
| Mobbs et al., 201628 | 2009–2013 | 0 | NA | NA | 105 | 4 | 3.8% | NA |
| Total (post-2008) | 127 | 3 | 2.4% | 201 | 5 | 2.5% | No (p = 1.0) | |
Data for this cohort were cross-referenced between this publication and that of Burkus et al., 2013.24
This study provided a demarcation in data, allowing for a separate analysis of pre-2008 and post-2008 cases. All other papers in which the data collection period spanned this demarcation were excluded.
Retroperitoneal Versus Transperitoneal Approach
A total of 1264 patients across 9 studies were included (Table 3). Three of these studies used both retroperitoneal and transperitoneal approaches17,22,23 whereas 6 expressly noted only retroperitoneal.11,12,19,21,25,26 Of those 3, only Sasso and colleagues observed a statistically significant increase in risk associated with a transperitoneal approach (13.3% vs 1.7%, p = 0.017).22 Pooled analysis demonstrated statistical significance, with 5 of the 58 transperitoneal approaches resulting in RE versus 38 of 1206 retroperitoneal approaches (8.6% vs 3.2%, p = 0.043).
Procedure Performed (ALIF vs TDR)
Thirteen studies were included in the risk analysis of ALIF with and/or without rhBMP-2 versus TDR (Table 4), all of which included ALIF (n = 1468)9,11,12,15–17,21–23,25–28 and 2 of which included TDR (n = 246).21,23 Across these 13 studies, the ALIF RE rates ranged from 0.0%15–17 to 9.8%,12 with a pooled average of 2.8%. Neither study using both techniques found statistical significance between groups. The TDR RE rates ranged from 2.0%23 to 9.8%,21 with a pooled average of 3.3%. The difference in pooled averages was not statistically significant (p = 0.84).
Single-Level Versus Multilevel Exposure
Across the 7 included studies (Table 5), the rate of RE with a single-level procedure ranged from 0.0%17 to 12.2%.21 Four of these studies also included multilevel procedures,11,12,21,29 with RE rates ranging from 4.1%11 to 13.3%.12 No included study or pooled analysis found statistical significance between groups. Of the 924 patients who underwent single-level procedures, 34 developed RE versus 14 of 282 patients who underwent multilevel surgery (3.7% vs 5.0%, p = 0.38).
Use of rhBMP-2
Thirteen studies that explicitly commented on rhBMP-2 versus bone graft use were included in this risk analysis (Table 6).9,11,12,15–17,21–23,25–28 RE rates in the rhBMP-2 cohorts ranged from 0.0%17 to 9.5%12 and rates in the control groups ranged from 0.0%17 to 10.0%.12 Combining these 13 studies, there was a significantly higher risk associated with rhBMP-2, with 29 of the 577 patients who received rhBMP-2 experiencing RE versus 20 of the 1137 patients in the control group (5.0% vs 1.8%, p < 0.001).
In order to assess the risk of RE associated with rhBMP-2 use prior to versus after the institution of rhBMP-2 handling precautions, we chose July 2008 when the FDA warning was issued as the demarcation (Table 7). In 7 studies all surgeries were performed prior to 2008,16,17,22,23,25–27 in 3 all surgeries were performed after 2008,9,15,28 and 1 study specified the number of patients in the experimental and control groups prior to and after this demarcation and thus was included as 2 separate cohorts.11 Although the increase in risk with rhBMP-2 was statistically significant prior to 2008 (5.3% vs 1.0%, p < 0.001), rates were comparable between rhBMP-2 and control cohorts after 2008 (2.4% vs 2.5%, p = 1.0).
Discussion
The anterior approach to the lumbar spine offers a unique biomechanical advantage, is associated with less blood loss and less manipulation of neural elements as opposed to other approaches to the lumbar spine, and preserves the paraspinal musculature and posterior ligaments if used as a stand-alone procedure. However, the anatomy encountered poses unique access-related complications.
Complication rates, particularly those associated with damage to intraperitoneal structures, have decreased with the advent of the retroperitoneal approach as opposed to a transperitoneal approach and the transition away from laparoscopic techniques, which limit direct visibility and are associated with a steep learning curve. Despite this, the retroperitoneal approach still poses unique risks such as damage to major vasculature or the hypogastric plexus, with resultant RE and infertility. As such, the risk of RE deters some surgeons from using an anterior approach in men with child-siring potential. However, the subjective nature of questionnaires developed for detection of RE as well as the scarcity of publications on the complications in men specifically has impeded an understanding of the true incidence and surgical risk factors of RE.
The reported incidence of RE varies greatly in the literature, from 0% in several of the studies examined to 45% (self-reported).30 Notably, the highest estimates tend to be associated with laparoscopic approaches. Indeed, the present systematic review of open and mini-open procedures found no incidence greater than 10%. Still, the included pooled data showed an incidence of only 2.3%, with nearly half of these cases resolving spontaneously (45.8%). Notably, this rate of resolution is likely to be an underestimation because "persistence" was diagnosed as early as 6 months,27 whereas resolution occurred as late as 48 months.22 Tepper and colleagues posited that RE may be overestimated by questionnaires, particularly in cases in which ongoing litigation may provide an incentive for patients to report more postoperative complications.12 However, although their study did show a gross overestimation by their questionnaire (41.7% self-reported decreased ejaculate volume as opposed to 9.8% RE detected in laboratory analysis), their quantitative data still detected the highest rate encountered in this systematic review. In alignment with their hypothesis, 39 of the 41 patients in their cohort were involved in workers’ compensation claims.
Surgical risk factors for RE have proven to be difficult to delineate in the absence of complete data stratified by sex. In their 2002 and 2003 publications on the same large multicenter randomized controlled trial (RCT), Burkus et al. and Sasso et al. found a statistically significant increase in RE following a transperitoneal versus a retroperitoneal approach (13.3% vs 1.7%, p = 0.017),22,31 a finding that was not replicated to statistical significance in subsequent RCTs by the same author group comparing ALIF with rhBMP-2 to TDR (4.2% vs 1.9%, p = 0.406)23 or ALIF with rhBMP-2 to ALIF with bone graft in which no cases of RE were reported.17 Combining the 9 studies for which data on approach stratified by sex could be determined, the incidence of RE was significantly higher, with a transperitoneal approach resulting in 5 of 58 males (8.6%) versus 38 of the 1206 males having some degree of RE after undergoing a retroperitoneal approach (3.2%) (p = 0.043). However, it should be noted that very few patients included in this analysis underwent a transperitoneal approach.
Because the present study sought to determine incidence and risk factors of RE in an anterior lumbar approach in general, TDR was not an exclusionary criterion. In both studies that included both ALIF and TDR, the RE rates were similar between groups.21,23 In pooled analysis of RE risk based on procedure performed, the difference in rates was not statistically significant, with 41 of the 1468 patients who underwent ALIF developing RE versus 8 of the 246 who underwent TDR (2.8% vs 3.3%, p = 0.84).
Because of the greater exposure required, multilevel procedures have been hypothesized as a risk factor. Of the 4 studies that included both single- and multilevel procedures and reported the number of men undergoing each, no studies found a significant difference in RE rates between the 2 groups.11,12,21,29 An additional 3 studies did not include multilevel procedures, but specifically stated the number of men with single-level procedures and thus were included.17,22,23 Across these 7 included studies, 34 of 924 men who had single-level procedures developed RE (3.7%) versus 14 of the 282 who underwent multiple-level procedures (5.0%). This difference was not statistically significant (p = 0.38).
Although RE has traditionally been thought of as an access-related complication, several studies support the hypothesis that in addition to direct damage to the hypogastric plexus, RE may also be caused by neuroinflammatory effects of rhBMP-2 on the plexus.10,11,24 In the aforementioned RCTs published both by Burkus and Sasso and colleagues in 2002 and 2003, a nonstatistically significant increase in RE rates was noted in patients who underwent ALIF with rhBMP-2 in comparison to ALIF with ICBG.22,31 Because this study was underpowered, the same group published a paper in 2013 combining the results of that trial with their subsequent RCTs comparing rhBMP-2 to ICBG or TDR.17,23,24,32 Although none of their other trials found a significantly increased risk of RE with rhBMP-2, their 2013 paper pooling the results found a nonstatistically significant increase in incidence (3.4% vs 1.7%, p = 0.242).24 However, the authors note that patients were counseled preoperatively on the risk of RE and subsequently self-reported any complications but were not specifically questioned on sexual function. Consequently, no cases of RE were reported in 3 of their 5 included trials.24
In a 2012 retrospective analysis of 10 years of data on consecutive ALIF cases, Comer and colleagues found a statistically significant increase in risk, with 15 of 239 (6.3%) patients in the ALIF with rhBMP-2 group developing RE versus 2 of 233 (0.9%) patients in the control group (p = 0.0012).11 Notably, of these 17 patients who reported RE, 16 cases were confirmed with laboratory analysis. One patient declined laboratory testing but was diagnosed by a urologist on clinical grounds. Whereas none of the other comparative studies included in the rhBMP-2 risk analysis showed an increased risk,12,17,21–23 pooled data showed a statistically significant increase, with 29 of 577 (5.0%) patients in the rhBMP-2 cohort developing RE compared to 20 of 1137 (1.8%) in the control group (p < 0.001).
Comer et al. suggest that some of the risk associated with rhBMP-2 may be modifiable by careful handling and delivery of the substance.11 They note that after the 2008 FDA-issued warning on its complications, they instituted measures to prevent inadvertent spread outside of the cage and found that their RE incidence was cut in half. This claim is supported by Malham et al. who identified only 1 case of RE in their single-arm prospective study of 67 men who underwent ALIF with rhBMP-2 between 2009 and 2012, with careful handling and a surgical barrier preventing local spread.9
To assess the credibility of this hypothesis, the present study analyzed the risk associated with rhBMP-2 separately in papers in which all surgeries were performed prior to 200816,17,22,23,25–27 and papers in which all surgeries were performed after 2008.9,15,28 Papers in which the data collection period spanned this demarcation were excluded, with the exception of 1 paper in which the authors reported on the number of patients in the experimental versus control group prior to and after 2008.11 Prior to 2008 there was a substantially higher risk of RE associated with rhBMP-2, with 20 of the 375 patients who received rhBMP-2 developing RE as opposed to 9 of the 875 in the control group (5.3% vs 1.0%, p < 0.001). After 2008 this significance was lost, with 3 of the 127 patients in the rhBMP-2 group versus 5 of the 201 patients in the control group (2.4% vs 2.5%, p = 1.0). There is a nonstatistically significant trend toward a decrease in risk of ALIF with rhBMP-2 before versus after 2008 (5.3% vs 2.4%, p = 0.22), suggesting that the FDA warning alerted surgeons to the detrimental effects of rhBMP-2 on soft tissue and that the resultant precautions mitigated the risk of RE.
The possibility must be entertained that this decrease may be a byproduct of scarcer reporting, less frequent use of BMP, or less frequent utilization of ALIF in men. It is difficult to sort out such confounders without more clarity on methods. However, the 2012 publication by Comer et al. speaks against these potential confounders.11 Throughout the study years, both direct questioning and questionnaire were used. Thus, frequency of reporting was unlikely to play a significant role. Furthermore, these authors specifically noted a clearly demarcated decrease in RE rates observed after the FDA warning. Although it is possible that this warning also directly caused them to use BMP or ALIF less frequently, they do note that they were still using BMP routinely during this period, but with additional precautions.
There are numerous limitations to the present study. What specifically constituted RE was rarely specified. Individual studies’ methods were difficult to discern and variable when specifically reported. A problem in much of the RE literature is the subjective nature of screening with questionnaires or direct questioning. Given the seriousness of this problem, it is conceivable that some of these men did undergo subsequent objective testing or saw specialists. However, subsequent workup or absence thereof is not documented in the literature. It is argued that subjective screening may underestimate RE because men may be hesitant to speak about sensitive issues with clinicians. Our results did demonstrate a higher rate of RE in the only study that used laboratory testing in all patients, although it should be noted that it was not specified whether objective laboratory detection met a threshold capable of revealing clinically significant infertility. Furthermore, in that study there was no follow-up laboratory analysis after 6 months, so delayed resolution may have been missed. Additionally, just as it is conjectured that men may be reluctant to discuss these issues, it may be reasonably argued that men who are not reporting these symptoms simply may not be bothered by them, given that RE is clinically significant only in the context of fertility. We believe that the significance of this overlooked discernment lies in proper patient selection.
Conclusions
From the combined data of 18 studies, there was a 2.3% incidence of RE in men undergoing first-time open anterior lumbar surgery, with nearly half of cases resolving spontaneously. Analysis of risk factors was limited by the scarcity of studies providing male-specific data, but the data that were able to be assessed showed an increased risk of RE associated with a transperitoneal approach and with the use of rhBMP-2. Given that the role of rhBMP-2 in RE is probably neuroinflammatory and that most of the data available were collected prior to the FDA warning that addressed its neuroinflammatory properties, it is feasible that these results overestimate the risk if care is taken to prevent spread of the drug outside the cage. This is reflected by the loss of statistical significance in the risk associated with rhBMP-2 if surgeries performed prior to 2008 are excluded from analysis. Furthermore, the inherent risk of RE in anterior lumbar surgery is probably overestimated as an artifact of the inclusion in meta-analyses of early data including the high-complication laparoscopic approaches, which are no longer widely used, and the disproportion of data demonstrating the artificially high rate of the true risk of rhBMP-2 as discussed here. Pooled analysis did not find a significant difference in the rates of RE in men undergoing single-level versus multilevel procedures or undergoing ALIF versus TDR. Ultimately, patients should be counseled on this potential complication and thoughtful consideration should be given to alternatives, but the relatively low incidence of RE and its tendency to self-resolve should not absolutely preclude open retroperitoneal anterior lumbar surgery or the judicious use of rhBMP-2 in men with child-siring potential. Future studies including the objective assessment of RE in men who elect to undergo anterior surgery may be beneficial.
Disclosures
The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
Author Contributions
Conception and design: Prestigiacomo, Body, Plummer, Krueger, Cheng. Acquisition of data: Body, Plummer, Krueger. Analysis and interpretation of data: Body, Plummer. Drafting the article: Body, Plummer, Krueger. Critically revising the article: Prestigiacomo, Plummer, Krueger, Virojanapa, Nasser, Cheng. Reviewed submitted version of manuscript: Prestigiacomo, Body. Approved the final version of the manuscript on behalf of all authors: Prestigiacomo. Statistical analysis: Body. Study supervision: Prestigiacomo, Virojanapa, Nasser, Cheng.
References
- 1↑
Heary RF, Yanni DS, Halim AY, Benzel EC. Anterior lumbar interbody fusion. In: Steinmetz MP, Benzel EC, eds.Benzel’s Spine Surgery. 4th ed. Elsevier;2017:655–666.
- 2↑
Muller W. Transperitoneale Freilegung der Wirbelsaule bei tuberkuloser spondylitis. Dtsch Z Chir. 1906;85:128.
- 4↑
Lane JD Jr, Moore ES Jr. Transperitoneal approach to the intervertebral disc in the lumbar area. Ann Surg. 1948;127(3):537–551.
- 5↑
Guyer RD, Otero-Lopez E. Lumbar total disc arthroplasty. In: Steinmetz MP, Benzel EC, eds.Benzel’s Spine Surgery. 4th ed. Elsevier;2017:1591–1596.
- 6↑
Dodwad SM, Mroz TE, Hsu WK. Biologics in spine fusion surgery. In: Steinmetz MP, Benzel EC, eds.Benzel’s Spine Surgery. 4th ed. Elsevier;2017:280–284.
- 7↑
Skovrlj B, Koehler SM, Anderson PA, et al. Association between BMP-2 and carcinogenicity. Spine (Phila Pa 1976). 2015;40(23):1862–1871.
- 8↑
Hindoyan K, Tilan J, Buser Z, et al. A retrospective analysis of complications associated with bone morphogenetic protein 2 in anterior lumbar interbody fusion. Global Spine J. 2017;7(2):148–153.
- 9↑
Malham GM, Parker RM, Ellis NJ, et al. Anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2: a prospective study of complications. J Neurosurg Spine. 2014;21(6):851–860.
- 10↑
Murray-Ortiz G, Park MS, Uribe JS. Anterior and lateral lumbar instrumentation. In: Winn HR, ed. Youmans and Winn Neurological Surgery. 7th ed. Elsevier;2017:2681–2686.
- 11↑
Comer GC, Smith MW, Hurwitz EL, et al. Retrograde ejaculation after anterior lumbar interbody fusion with and without bone morphogenetic protein-2 augmentation: a 10-year cohort controlled study. Spine J. 2012;12(10):881–890.
- 12↑
Tepper G, Rabbani R, Yousefzadeh M, Prince D. Quantitative assessment of retrograde ejaculation using semen analysis, comparison with a standardized qualitative questionnaire, and investigating the impact of rhBMP-2. Spine (Phila Pa 1976). 2013;38(10):841–845.
- 13↑
Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009;151(4):264–269,W64.
- 14↑
Wright JG, Swiontkowski MF, Heckman JD. Introducing levels of evidence to the journal. J Bone Joint Surg Am. 2003;85(1):1–3.
- 15↑
Amaral R, Ferreira R, Marchi L, et al. Stand-alone anterior lumbar interbody fusion – complications and perioperative results. Rev Bras Ortop. 2017;52(5):569–574.
- 16↑
Arlet V, Jiang L, Steffen T, et al. Harvesting local cylinder autograft from adjacent vertebral body for anterior lumbar interbody fusion: surgical technique, operative feasibility and preliminary clinical results. Eur Spine J. 2006;15(9):1352–1359.
- 17↑
Burkus JK, Sandhu HS, Gornet MF, Longley MC. Use of rhBMP-2 in combination with structural cortical allografts: clinical and radiographic outcomes in anterior lumbar spinal surgery. J Bone Joint Surg Am. 2005;87(6):1205–1212.
- 18↑
Edgard-Rosa G, Geneste G, Nègre G, Marnay T. Midline anterior approach from the right side to the lumbar spine for interbody fusion and total disc replacement: a new mobilization technique of the vena cava. Spine (Phila Pa 1976). 2012;37(9):E562–E569.
- 19↑
Kang BU, Choi WC, Lee SH, et al. An analysis of general surgery-related complications in a series of 412 minilaparotomic anterior lumbosacral procedures. J Neurosurg Spine. 2009;10(1):60–65.
- 20↑
Quraishi NA, Konig M, Booker SJ, et al. Access related complications in anterior lumbar surgery performed by spinal surgeons. Eur Spine J. 2013;22(1 Suppl):S16–S20.
- 21↑
Lindley EM, McBeth ZL, Henry SE, et al. Retrograde ejaculation after anterior lumbar spine surgery. Spine (Phila Pa 1976). 2012;37(20):1785–1789.
- 22↑
Sasso RC, Kenneth Burkus J, LeHuec JC. Retrograde ejaculation after anterior lumbar interbody fusion: transperitoneal versus retroperitoneal exposure. Spine (Phila Pa 1976). 2003;28(10):1023–1026.
- 23↑
Gornet MF, Burkus JK, Dryer RF, Peloza JH. Lumbar disc arthroplasty with Maverick disc versus stand-alone interbody fusion: a prospective, randomized, controlled, multicenter investigational device exemption trial. Spine (Phila Pa 1976). 2011;36(25):E1600–E1611.
- 24↑
Burkus JK, Dryer RF, Peloza JH. Retrograde ejaculation following single-level anterior lumbar surgery with or without recombinant human bone morphogenetic protein-2 in 5 randomized controlled trials: clinical article. J Neurosurg Spine. 2013;18(2):112–121.
- 25↑
Loguidice VA, Johnson RG, Guyer RD, et al. Anterior lumbar interbody fusion. Spine (Phila Pa 1976). 1988;13(3):366–369.
- 26↑
Newman MH, Grinstead GL. Anterior lumbar interbody fusion for internal disc disruption. Spine (Phila Pa 1976). 1992;17(7):831–833.
- 27↑
Brau SA. Mini-open approach to the spine for anterior lumbar interbody fusion: description of the procedure, results and complications. Spine J. 2002;2(3):216–223.
- 28↑
Mobbs RJ, Phan K, Daly D, et al. Approach-related complications of anterior lumbar interbody fusion: results of a combined spine and vascular surgical team. Global Spine J. 2016;6(2):147–154.
- 29↑
Jarrett CD, Heller JG, Tsai L. Anterior exposure of the lumbar spine with and without an “access surgeon”: morbidity analysis of 265 consecutive cases. J Spinal Disord Tech. 2009;22(8):559–564.
- 30↑
Kaiser MG, Haid RW Jr, Subach BR, et al. Comparison of the mini-open versus laparoscopic approach for anterior lumbar interbody fusion: a retrospective review. Neurosurgery. 2002;51(1):97–105.
- 31↑
Burkus JK, Gornet MF, Dickman CA, Zdeblick TA. Anterior lumbar interbody fusion using rhBMP-2 with tapered interbody cages. J Spinal Disord Tech. 2002;15(5):337–349.
- 32↑
Burkus JK, Transfeldt EE, Kitchel SH, et al. Clinical and radiographic outcomes of anterior lumbar interbody fusion using recombinant human bone morphogenetic protein-2. Spine. (Phila Pa 1976).2002;27(21):2396–2408.
- 33↑
Fischer CR, Braaksma B, Peters A, et al. Outcomes and complications of the midline anterior approach 3 years after lumbar spine surgery. Adv Orthop. 2014;2014:142604.
