22.2.1 Maintenance of Radiographic Correction

Research on the outcomes of surgical correction of AIS has long been focused on maintenance of the deformity correction, but this research has typically been limited to reporting findings only in the coronal plane. ^{1 , 2 , 3 , 4} Recent long-term outcome studies have reported modest average loss of coronal correction in primary thoracic curves treated with both anterior and posterior approaches (range of 2- to 9-degree loss of thoracic correction) at an average of ≥15 years postoperatively. ^{5 , 6 , 7 , 8} Two of these studies reported a minimal increase in the unfused lumbar curve (2- to 3-degree increase). ^{5 , 6} Similarly, recent reports of primary thoracolumbar curves at long-term follow-up report minimal average loss of correction from early postoperative (3- to 6-degree increase in Cobb angle) and maintenance of thoracic correction. ^{9 , 10 , 11} Studies that have included both curve types also report low average loss of coronal correction of the primary Cobb angle in long-term follow-up studies. ^{12 , 13}

Many of the earlier studies only focused on coronal Cobb angle as an outcome, but the understanding of scoliosis as a 3D deformity has grown, and emphasis on evaluating more than just the coronal plane has become more commonplace. A few recent studies looking at long-term outcomes have included radiographic findings in both the sagittal and axial planes. Larson et al found similar sagittal plane outcomes in primary thoracic curves treated with either selective thoracic fusion or long fusion including the lumbar curve. ⁶ Two studies examining the long-term outcomes of primary thoracolumbar/lumbar curves treated with anterior instrumentation reported maintenance of sagittal thoracic kyphosis and lordosis from the immediate postoperative to the final follow-up visit. ^{9 , 10} One recent study on the outcomes of thoracolumbar curves after dual-rod anterior instrumentation reported maintenance of thoracic kyphosis but found a significant increase in overall lumbar lordosis and also the lordosis below the fused segments. ¹¹ Only two studies have examined the axial plane, both utilizing 2D methods (Perdriolle and Nash–Moe methods) and found no change in apical vertebral rotation between early and late postoperative visits in thoracolumbar curves. ^{9 , 11}

With the advent of biplanar simultaneous radiograph acquisition, advanced methods of evaluating the 3D deformity associated with scoliosis have been developed. ^{14 , 15 , 16} A better method to assess axial rotation and a more accurate sense of the sagittal deformity (particularly in the thoracic spine) has been identified and validated with this new technology (refer to Chapter 9 for more details). Future research on long-term outcomes will need to be focused on the inclusion of these new methods of measuring the deformity in order to truly understand whether the correction is maintained in all three planes at ≥15 years of follow-up (Fig. 22‑4).

It is unclear if studies that have evaluated average correction over the duration of the postoperative course similarly represent the clinical outcomes. Is an average loss of 9 degrees of coronal Cobb angle acceptable to the surgeon and the patient? At what point does the loss of coronal correction warrant revision surgery? Unknown among these averages are patients whose loss of correction may be in an unacceptable range and how their quality of life (QOL) suffers differently from the mean. Attempts at correlating QOL measures with the final Cobb angle have yielded insignificant relationships. ^{17 , 18 , 19} However, as much as “poor outcomes” get lost among averages, these same poor outcomes may not be identifiable in a statistical analysis that attempts to identify linear relationships within correlation analyses. The aforementioned sagittal and axial measures are also critical components needed to factor into the algorithm of understanding what the range of “optimal” radiographic outcome is in the long term. Research efforts need to be invested to help the patient understand what these averages of radiographic measures suggest. What is their risk of having a long-term radiographic outcome that will interfere with their activities of daily living and/or prompt discussion from their surgeon when considering a revision surgery? What does radiographic progression mean to the patient and what questions are still unanswered? Could a 10-degree loss of coronal correction be an indicator of another adverse event that would affect patient outcomes like an indolent infection or a nonunion? These factors require further investigation in future long-term outcome studies.

A better understanding of the relationship between radiographic outcomes and health-related quality-of-life (HRQOL) assessment is warranted. When evaluating patients preoperatively, there is minimal correlation between radiographic status and HRQOL, ²⁰ which indicates that variables other than radiographic alignment/measures of the deformity must also affect these outcome scores. When assessing change between preoperative and 2-year outcomes, the radiographic assessment shows a significant improvement; however, little correlation between the radiographic assessment and the HRQOL questionnaire score exists. ²¹ This needs to be assessed in the long term. One study evaluated the longer-term outcomes of 109 patients with greater than 10 years of follow-up and found that patients who lost ≥10 degrees of correction had lower HRQOL scores. ²² Sucato et al found that coronal balance was the only radiographic parameter that correlated to improved appearance and satisfaction scores in a cohort of AIS patients greater than 10 years postoperatively. ²³

Recent minimum 10-year data of primary thoracic curves from the Harms Study Group have shown overall maintenance of the radiographic correction, with a 4-degree average increase in thoracic Cobb angle from initial postoperative to 10-year follow-up. Coronal balance improved during the postoperative course and sagittal kyphosis remained increased from pre-op at the 10-year mark. Of the 25 patients (14%) who had a loss of correction of ≥10 degrees, less than 2% reported a 10-year Scoliosis Research Society-22 (SRS-22) pain score that was greater than 2 standard deviations below the mean for similar aged controls.s. Literatur

Additionally, Lonner et al ²⁵ examined disc degeneration at the caudal segments below a spinal fusion 10 years later. They found that disc wedging subjacent to the lower instrumented vertebra (LIV; >5) and LIV translation (>2 cm) led to a sixfold increase in significant disc degeneration. The severity of disc degeneration was not associated with the number of levels fused (p = 0.2), the surgical approach (p = 0.8), or the construct type (p = 0.3). No significant association was established between 10-year composite radiographic score (CRS) and SRS-22 scores. These recent data imply that, while disc wedging and translation resulted in a substantial increase in disc degeneration, functional limitations could not be established at 10 years following surgery. Longer-term data are certainly required.

22.2.2 Cosmesis

The outward deformity of the trunk associated with scoliosis and/or the asymmetrical shoulder height are typically the first signs of scoliosis noticed by the patient, their family, or the provider referring them for treatment. The cosmetic deformity is often a significant concern of the patient, and improvement of this external deformity may be a major expectation associated with the decision to undergo surgery (Fig. 22‑5).

Although limiting the progression of scoliosis over time may be one of the primary goals for the surgeon, obtaining and maintaining correction of the appearance of trunk deformity over time is likely more important to the patient. Yet, the inclusion of trunk shape measures in long-term (≥10 years postoperative) outcomes studies is limited (Fig. 22‑6).

Larson et al completed the most comprehensive assessment of trunk shape measures. Within their cohort of patients who underwent selective thoracic fusion, those with coronal imbalance greater than 2 cm actually improved from 43% of the cohort at intermediate follow-up (5.5 years) to 29% at a long-term average of 20.7 years of follow-up. ⁶ A trunk shift (midpoint of the chest relative to the pelvis) greater than 2 cm remained stable over the same time period. They compared shoulder height, coronal balance, trunk shift, and scoliometer readings at long-term follow-up visits between patients undergoing selective thoracic fusion and those undergoing long thoracic plus lumbar fusion and reported no significant differences.

Helenius et al compared the long-term follow-up analysis of Harrington versus Cotrel–Dubousset (CD) instrumentation and found at least half of each cohort had waist asymmetry and the average rib prominence in both groups was 13 and 11 degrees, respectively. ²⁴ There was no suggestion of whether this had remained stable, improved, or worsened over the postoperative course. Sudo et al reported stable coronal balance over an average of 15 years postoperatively in patients with thoracic curves treated with anterior spinal fusion. ⁸ Delfino et al reported a significant improvement from an early postoperative average coronal balance of 1.6 cm to long-term follow-up average of 0.8 cm in selective anterior thoracolumbar fusions. ⁹

It is unclear from the existing literature whether trunk deformity correction is maintained and/or found to be acceptable by patients at long-term follow-ups into mid to late adulthood. Conducting these long-term studies in a retrospective fashion likely limits the type of data available from the medical record in order to be able to assess changes that exist from the early postoperative period to the long-term follow-ups. Prospective studies are required in order to obtain comprehensive trunk shape data available at relevant time periods over the course of treatment. Frustration may also exist among researchers in regard to the existing methods for assessing cosmesis, as studies suggest that current methods show little to no correlation with patients’ reported QOL scores ^{25 , 26 , 27} (Fig. 22‑7).

Despite this lack of correlation, understanding the magnitude of the deformity on clinical presentation is essential in providing a complete picture of treatment outcomes. Prospective studies collecting long-term follow-up data with consistent and standardized measures of trunk shape data may determine the best method to assess changes in trunk shape over the long term, how these changes are perceived by the patient, and how those perceptions affect the patient’s QOL. The Harms Study Group’s prospective registry study of AIS was established in 1995, and longer-term follow-up data continue to be collected.

22.2.3 Pulmonary Function

Much interest revolves around the outcome of pulmonary function in patients with AIS, as there have been studies suggesting increased morbidity and mortality related to reduced pulmonary function in a progressive untreated scoliotic deformity. ^{28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36} One of the major goals associated with halting curve progression via surgical correction is to prevent pulmonary function decline and potentially restore any existing deficits. Two-year to mid-term (5-year) follow-up studies have suggested relatively stable percent predicted pulmonary function values based on pulmonary function testing (PFT). However, an open anterior approach and thoracotomy lead to statistically significant decreases in average PFT values. ^{31 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44} The clinical significance of these decreases, implications for the duration of the patient’s life, and whether the surgical intervention is better than the long-term natural history remain unanswered from these early follow-up studies. A few authors have examined the pulmonary function of AIS patients who are greater than 15 years postoperative. Two studies by the same authors that looked at long-term outcomes of primary thoracic (Lenke type 1) curves with an average follow-up of 15 years demonstrated significant declines in pulmonary function following anterior spinal fusion. ^{8 , 45} The forced vital capacity (FVC) values dropped by 7 to 8% in both studies, and the forced expiratory volume in 1 second (FEV1) dropped an average of 11%. ^{8 , 45} However, none of the patients reported pulmonary-related symptoms or functional issues. Ruiz-Juretschke et al studied 24 thoracolumbar/lumbar curves that had undergone anterior spinal fusion at a range of 12 to 29 years postoperatively and reported that the percent predicted values for FVC and FEV1 were within normal ranges. ⁴⁶ In a subset of patients, they found no significant difference in absolute FVC and FEV1 values from preoperative to the long-term postoperative visit. These findings seem to reflect those of the early outcome studies, whereby patients who underwent an anterior open fusion for thoracic curves seem to suffer a modest decline in the percent predicted values.

Pehrsson et al examined the pulmonary function of 141 operative scoliosis patients within 2 years of surgery and again at least 20 years postoperatively. ³² All patients underwent posterior fusion with Harrington instrumentation. All patients experienced significant improvements in vital capacity (VC), FEV1, and total lung capacity (TLC) from pre- to postsurgery and from postsurgery to long-term follow-up regardless of curve type. Eight percent of patients had a VC that was less than 70% predicted, but the authors noted that an unmentioned number of patients had other possible medical explanations for the reduced values. These authors also compared the average PFT values of patients with curves treated by bracing (less severe deformity) to normal controls. Although the average long-term follow-up percent predicted values were significantly lower in the operative patients than both normal controls and previously braced patients, the improvement from pretreatment to long-term follow-up was significantly greater in the operative than in the braced patients. Although this is not quite a direct comparison to the natural history of the operative magnitude of untreated curves, this study suggests that operative intervention provided a substantial increase in pulmonary function but did not restore function to the level of patients with less severe deformity or normal controls (Fig. 22‑8).

Some studies have evaluated the mortality associated with untreated idiopathic scoliosis at a follow-up of greater than 30 years following diagnosis. ^{30 , 34 , 35} The rate of mortality possibly due to respiratory compromise secondary to scoliosis ranged from 1/219 (0.5%) to 5/117 (4.3%). Understanding the morbidity and mortality of operatively treated patients at this late stage of follow-up (beyond 30 years postoperative) and whether the natural history is altered has yet to be determined. Since the rate of mortality is highly variable and in reality likely quite low for lesser magnitudes of deformity, it will be challenging to power a study to adequately identify a significant reduction in mortality with operative intervention. However, research into understanding how the deformity itself and the choice to surgically correct the deformity impact the ability to participate in life-long activities requiring adequate pulmonary function (e.g., sports participation, dance, keeping up with running after a toddler) is critical, as some of the improvements seen from pre- to postoperative in pulmonary function may (or may not) translate into better QOL due to increased cardiopulmonary efficiency.

22.2.4 Spine Function, Mobility, and Health

Degeneration of the spine in aging adults is a natural phenomenon, and surgical treatment for this condition has grown quite significantly, with one study reporting a 2.4-fold increase in surgical treatment of lumbar degenerative disc disease (DDD) in the United States during 2000 to 2009. ⁴⁷ Concern over whether scoliotic deformity, and/or its treatment, can increase the likelihood of developing DDD is a valid question for surgeons and patients. Attempts have been made to determine whether spinal fusion in adolescence for idiopathic scoliosis increases the incidence of developing signs of disc degeneration, particularly below the fusion. Despite numerous studies prior to the 2000s attempting to answer this question, the debate still continues into the new millennium with many researchers continuing to search for a clear answer. ^{1 , 2 , 35 , 48}

At least four long-term studies since 2000 have found increased rates of radiographic evidence of DDD in postoperative AIS patients compared to normal controls. ^{12 , 49 , 50 , 51} In two studies from Sweden, surgically treated AIS patients demonstrated a 24% rate of degenerative disk (DD) changes compared to 0% in their group of controls. ^{12 , 50 , 51} One of their studies identified a correlation with increased lumbar pain and DD at the lowest unfused disc. ⁵⁰ A 2018 study from Japan found increased Modic changes in postoperative AIS patients (58%) compared to controls (14%), but no difference at the level of severity warranting diagnosis of lumbar DDD (Pfirrmann grade 4 to 5; 61.5% in AIS patients vs. 65.5% in controls) was found. ⁴⁹ They found that patients with Modic changes had increased lumbar curve magnitudes and increased kyphosis at the thoracolumbar junction (T10–L2) compared to patients with no Modic changes. These patients also had lower mental health scores on the SRS-22 questionnaire.

Although these studies suggest some increased risk of developing radiographic spinal changes that could lead to pain, disability, and need for further surgical intervention, it is unclear from these comparisons whether changes were caused by the scoliotic deformity or the surgical treatment of the deformity. One of the studies from Sweden evaluated a cohort of patients with a less severe deformity who had been treated by bracing in addition to long-term postoperative patients and found that they also had increased radiologic findings compared to normal controls (16 vs. 0%) but no significant difference from the surgically treated cohort. ¹⁷ Another paper by Akazawa et al evaluated 25 patients with AIS who had not undergone surgery and had reached middle age. ⁵³ Similar to the findings with postoperative AIS patients, the nonoperative AIS cohort demonstrated a higher rate of Modic changes (56%) compared to controls (8%). The rate of Pfirrmann grade 4 or 5 was 80% for the nonoperative AIS patients compared to 60% in controls. This finding was not significant (p = 0.059), and the authors did not comment on the 80% rate in nonoperative patients as compared to the 61.5% rate found in their previous study of postoperative AIS patients. The findings from these two studies suggest the scoliosis deformity itself results in increased spinal changes later in life and not the surgical management of the deformity, but this has yet to be confirmed by a direct study.

It has been suggested that fusions extending distally into the lumbar spine may increase the rate of disc degeneration. This suggestion does seem to be supported by two recent investigations. ^{51 , 52} Akazawa et al ⁵² reported positive rates for DDD (Pfirrmann grade 4 or 5) in 59% of patients fused to L3 or proximal compared to 100% in patients fused to L4 or distal. They also noted decreased lordosis and sagittal vertical axis in the patients with longer fusion. Danielsson and Nachemson ⁵¹ reported degenerative lumbar changes in 45% of patients fused to L5 compared to only 16% in those with fusions ending at L4 or higher; however, they did not find a difference between fusions ending at L3/above compared to L4/below.

One of the theories why scoliosis or surgical fusion performed to correct the scoliosis might lead to an increased risk of disc degeneration may be due to the potential for decreased normal spinal motion or motion that is not evenly distributed across the spine. Helenius et al ²⁴ studied the long-term outcomes of Harrington compared to CD instrumentation and had the patients perform functional tests, measures of spinal mobility, and trunk muscle strength. Squatting score, lumbar extension, and trunk side bending were significantly more impacted in the Harrington group. No correlations between the residual deformity of the thoracic and lumbar curves and the function, mobility, or strength measures were found. The authors noted that 24% of the Harrington patients had narrowing of the disc spaces and 22% had degenerative lumbar spondylosis, compared to none of these changes noted in the CD group. Green et al ⁵⁴ described long-term imaging in patients with greater than 10 years of postoperative follow-up and found that magnetic resonance imaging (MRI) demonstrated new disc pathology in 85% of patients; and, against theories of adjacent disc disease, most of the pathology was seen remote from the distal fusion level and was at the L5–S1 disc space.

Danielsson et al ⁵⁵ evaluated muscle endurance and spinal motion in patients at least 20 years after surgery or brace treatment. Both brace-treated and surgically treated scoliosis patients had decreased mobility and endurance compared to controls. Surgically treated patients had less motion than brace-treated patients, but no difference in endurance was observed. Lumbar motion was impacted in patients fused to L4 versus those fused to L3 or proximal. Similarly, patients fused to L1 or proximal had better lumbar motion than those fused to L2 or below. Better mobility and endurance was associated with increased function scores on the 36-Item Short-Form Survey (SF-36), and increased muscle endurance was associated with a lower score on the Oswestry Disability Index (ODI). The authors noted the lower values associated with these significant associations. They did not evaluate evidence of DDD related to the decreased mobility and endurance.

Understanding the intervertebral segmental motion changes in the distal segments following posterior spinal fusion has been a focus of the Harms Study Group. In these research efforts, it was revealed that preservation of vertebral motion segments (i.e., selective thoracic fusions) allowed greater distribution of functional motion across more levels and, with more distal instrumented vertebrae, there was significantly greater L2–L5 segment motion. ⁵⁶ In an effort to begin to understand the changes in mobility below the fusion over time, motion below the fusion in patients at their early and midterm postoperative follow-up time points was compared to patients with longer (>10 years) follow-up. The effect of time was not identified as the magnitude of distal motion observed in patients from early to longer follow-ups after surgery was similar ⁵⁷ (Fig. 22‑9).

None of these studies mention a need for conservative or operative intervention for any of these degenerative findings, so the clinical implications of these higher rates of radiographic degeneration are not clear. One study by Delfino et al ⁹ reported that 1 of their 35 subjects (2.9%) presented with symptomatic adjacent segment disc degeneration requiring surgical treatment. More research into understanding the spinal mobility of postoperative patients and the rate of degenerative disease that requires treatment and whether this rate is increased from a control population is needed.

22.2.5 Health-Related Quality of Life

Understanding how the disease process impacts patients’ lives and how the treatment affects their ability to resume a relatively disease-free life after treatment has been of interest for many years. The impetus behind developing HRQOL instruments may have been related to the need to audit or monitor the quality of care provided by physicians and care providers. Rosser ⁵⁸ traces this history back to 450 BC in Egypt, where Herodotus conducted the first recorded audit of medical services. The term QOL can be traced to the 1960s, ^{59 , 60} and HRQOL was first penned in the 1980s by health psychologists. ^{61 , 62} The first known instrument to attempt to assess and quantify a patient’s perception of his or her QOL was the Health Status Index, developed in the 1970s. ⁶³ Then, researchers started to develop disease-specific HRQOL instruments. Many tools are now available for functional impairment and disability, but the most widely utilized disease-specific tool for AIS is the Scoliosis Research Society (SRS) outcomes tool, which was first reported in the literature in 1999. ⁶⁴

The history of HRQOL development is critical in understanding the current state of long-term HRQOL research in surgically treated AIS patients. Since the most widely utilized outcomes tool in AIS was not developed until 1999, there are no preoperative or immediate postoperative data with this tool for comparison prior to this time. Many long-term studies conducted at this point in time utilize either rudimentary assessment of patients’ HRQOL perception or they only have scores available at the long-term follow-up visits. A number of studies of operative AIS patients have been able to utilize normal controls to compare validated HRQOL scores at the long-term follow-up visits. All but one found significant differences on some of the instruments and domains measured; however, there is no consistency across studies for which domains were different from controls. Padua et al examined 70 patients 15 to 28 years after Harrington instrumentation for AIS utilizing the SF-36 and the Roland and Morris Disability Questionnaire (RDQ). ¹⁸ The AIS postoperative cohort scored significantly less on physical function than normative data but significantly increased their average mental health score. Götze et al ¹⁹ studied 82 patients at an average of 16.7 years after Harrington instrumentation. The physical component of the SF-36 did not differ from age-matched controls, but the mental health component demonstrated significantly lower vitality, general mental health, and role activities because of emotional problems. Seventy-nine percent of patients reported no pain on the RDQ. Akazawa et al ⁶⁵ assessed 66 AIS patients at an average of 31.5 years postoperatively (21–41 years). The SRS-22, the RDQ, and a study-specific tool were utilized to assess the AIS patients and a group of age-, gender-, and BMI-matched controls. Function and self-image were lower in the AIS group, but no differences in pain or mental health were noted. The RDQ scores were also statistically similar. Larson et al ⁶ studied 28 patients with Lenke type 1B, 1C, or 3C curves at an average of 20 years postoperatively and found no difference from controls on SF-12 scores. Beyond the lack of consistency in which domains were different from controls across the studies, there is also little understanding from these findings as to what these statistical differences in average scores signify.

Two Swedish studies examined patients at least 20 years after surgery and not only utilized validated HRQOL scores compared to controls but also queried both surgical patients and controls on how the status of their back impacted activities of daily living. ^{17 , 51} In one study, surgical patients demonstrated significantly lower physical function scores using the SF-36 and also had significantly different ODI pain scores. Almost 50% of the surgical patients reported that they limited social activities because of their back compared to 15% of controls. The top two reasons surgical patients gave for limiting social activities were their belief that they could not physically participate in activities and fear of injury. The surgical patients also reported significantly more sick leaves compared to controls. ¹⁷ A follow-up study 2 years later confirmed these findings and also reported that the surgical patients reported significantly more pain in the lumbar region and buttocks/legs. However, the differences between the neck and thoracic region were not significantly different. ⁵¹

It is not surprising that patients with a scoliotic deformity requiring surgical correction would report differences from “healthy” controls at any point in their disease course. The most useful information to guide patients in treatment decision-making at the time of diagnosis (during adolescence) would be an understanding of how HRQOL is impacted in the long term between surgical and nonsurgical patients. There are currently no available studies that compare validated HRQOL scores in patients with similar magnitude deformities that chose surgical versus conservative management at greater than 15 years after diagnosis. In the study by Danielsson et al, ¹⁷ patients treated with bracing (less severe deformity at the time of diagnosis) were included, and HRQOL scores at least 20 years later were compared to a surgically treated cohort. This single study suggests that perhaps the HRQOL differences observed in the long term between postoperative AIS patients and controls is possibly due to the deformity and not the treatment. However, more studies will be necessary to confirm this hypothesis.

Another important question that plays a part in the surgeon’s and family’s decision-making at the time of surgery is the extent of fusion. Understanding how long fusions impact the HRQOL of the patient later in life is critical in helping educate these decisions for some curves (e.g., deciding whether to fuse both the thoracic and lumbar curves, or deciding whether to include L4 in the fusion or stop at L3). A few studies have evaluated the relationship between HRQOL scores and the level of LIV. Lavelle et al ⁶⁶ found no difference based on LIV in 22 patients for SRS-22, visual analog scale (VAS) pain, SF-36, or the ODI at 15 to 25 years after surgical correction with CD instrumentation. Götze et al ¹⁹ similarly did not find differences on the SF-36 or RDQ based on LIV in their series of 82 patients. However, Padua et al ¹⁸ did find that longer fusions were associated with more bodily pain, lower general health, and decreased vitality. Larson et al ⁶ found that patients with long fusion had slightly higher ODI scores (10 of 100) compared to short fusion ODI scores (4 of 100), but this did not reach significance. In their study, patients with long fusion reported better self-image and function after surgery on the SRS-24 questionnaire. Danielsson et al ⁵¹ reported no difference in pain or function based on whether a patient was fused at L3 or L4. Based on these conflicting findings, there is still much work to be done to understand the impact of fusion length on long-term HRQOL. Targeted large-volume multicenter studies on specific curve types will likely be required to truly understand the impact of fusion level choice on long-term HRQOL outcomes.

Identifying optimal specific HRQOL tools for AIS patients that can assess the impact of their disease over their lifespan is crucial to understanding how their treatment affects their long-term outcomes. Elucidating parameters such as minimal clinically important differences (MCID) and the minimum detectable change (MDC) for HRQOL instruments being used is essential, so that statistical differences in averages can be interpreted in a meaningful way. Mid-to late-life HRQOL comparisons with MCID and MDC between surgically treated patients and conservatively treated patients with similar deformities at diagnosis are necessary in order to help answer parents’ and patients’ questions concerning what the decision to have surgery means for their function as they age.

22.2.6 Pregnancy and Childbirth

With the predominant proportion of those impacted by AIS in the operative range of deformity being female prior to childbearing years, it is no surprise that future effects of scoliosis on pregnancy and childbirth would be a major concern to families. Understandably, discussion about surgical treatment raises many concerns for adolescent patients and their parents in terms of potential issues during pregnancy due to the instrumentation/fusion and potential for regional (spinal or epidural) anesthetic use during childbirth. ^{67 , 68} Concerns about curve progression during pregnancy are also a worry for both patient and surgeon, particularly in the case of selective thoracic fusion.

Many studies on this topic do not separate outcomes based on previous treatment type or baseline curve magnitude (mixed cohort of operative and conservatively managed patients). Some investigators have conducted long-term follow-up studies with the specific intent of understanding whether operative AIS treatment has negatively impacted the rate of childbirth and/or complicated the childbirth experience. The studies from Sweden ^{17 , 69} report a comparable number of childbirths in the postoperative AIS population compared to controls and brace-treated patients. They reported significantly more vacuum extractions in the postoperative scoliosis group and significantly increased self-reports that back conditions limited sexual activity for both scoliosis populations compared to controls. There was no correlation with curve progression after pregnancy and the number or age at first pregnancy. Investigators from Finland found increased cesarean section deliveries (23%) in patients who had been treated with Harrington rods compared to the general population (15%), but found no impact on curve progression. There were similar rates of back pain compared to rates of nonscoliotic patients in the literature. ⁷⁰ However, an Israeli study ⁷¹ found increased back pain requiring hospitalization during pregnancy (35%) in postoperative AIS patients compared to controls (0%) and greater rates of postpartum back pain (76%) compared to controls. Lavelle et al ⁷² studied a small group of patients with mixed diagnosis who had previously undergone anterior spinal fusion and found no issues of infertility and a cesarean section rate comparable to the national average but with only a 10% rate of epidural placement during delivery (compared to ~50% national average at the time).

A comprehensive review of the literature on pregnancy within AIS patients conducted by Dewan et al was published in 2018. ⁷³ This review included 22 articles with various time periods of follow-up and patients who had undergone a mix of treatment methods; thus, it is not clear exactly how surgical intervention would impact these findings. In their conclusion, they recommend conducting research with a higher level of evidence and including curve classification and surgical approaches as covariates in the analysis of outcomes. Their review was structured to answer three core questions: (1) How does scoliosis affect the success and timing of childbearing? (2) What happens to the spine as a result of pregnancy? (3) What are the obstetric and anesthetic considerations relevant in patients with AIS? In regard to the first question, they concluded that AIS patients had similar numbers of children at the same age as the general population; however, they also concluded that AIS patients may be less likely to become pregnant and require infertility treatment. For the second question, they concluded that AIS patients may expect higher rates of back pain during pregnancy and a statistically but clinically questionable progression of their curvature. Finally, the authors summarized that patients with instrumentation can have successful spinal anesthesia, but the number of attempts and minor complications may be higher than that in controls. In addition, they reported that the rate of cesarean section or perinatal complications was not increased in AIS patients.

As suggested by Dewan et al, higher levels of evidence are needed with larger sample sizes to subanalyze the outcomes based on curve type and surgical approach. ⁷³ The discrepancy of outcomes found in the existing studies may be due to the inclusion of a wide variety of deformity types and a mixed group of not just surgical approaches but also the extent of the fusions. The Harms Study Group has evaluated patients’ experiences with pregnancy and childbirth following surgical correction in the prospective AIS database registry. This has been prepared with the administration of an an obstetric and gynecologist designed questionnaire. The group found the following: there is no increased risk of prematurity (compared to national averages), patients who have had scoliosis surgery are less likely to get an epidural or spinal anesthesia, and there is a higher rate of C-section. Abdominal binders are reported to be helpful with back pain. Recommendations to patients from this research effort stated that patients should be sure to keep a copy of the X-rays to show their anesthesiologist in preparation for childbirth so they are aware of the lowest level of spinal fusion in order to minimize concerns for performing spinal anesthesia. ⁷⁴