Original Research

ABSTRACT

The importance of cost control in total knee arthroplasty is increasing in the United States secondary to both changing economic realities of healthcare and the increasing prevalence of joint replacement.

Surgeons play a critical role in cost containment and may soon be incentivized to make cost-effective decisions under proposed gainsharing programs. The purpose of this study is to examine the cost-effectiveness of all-polyethylene tibial (APT) components and determine what difference in revision rate would make modular metal-backed tibial (MBT) implants a more cost-effective intervention.

Markov models were constructed using variable implant failure rates and previously published probabilities. Cost data were obtained from both our institution and published United States implant list prices, and modeled with a 3.0% discount rate. The decision tree was continued over a 20-year timeframe.

Using our institutional cost data and model assumptions with a 1.0% annual failure rate for MBT components, an annual failure rate of 1.6% for APT components would be required to achieve equivalency in cost. Over a 20-year period, a failure rate of >27% for the APT component would be necessary to achieve equivalent cost compared with the proposed failure rate of 18% with MBT components. A sensitivity analysis was performed with different assumptions for MBT annual failure rates.

Given our assumptions, the APT component is cost-saving if the excess cumulative revision rate increases by <9% in 20 years compared with that of the MBT implant. Surgeons, payers, and hospitals should consider this approach when evaluating implants. Consideration should also be given to the decreased utility associated with revision surgery.

Continue to: All-polythylene tibial implants...

 

All-polyethylene tibial (APT) implants have been available for use in total knee arthroplasty (TKA) for decades. Except for one particular implant design, APT implants have shown equivalent functional outcome and survivorship to metal-backed tibial (MBT) components.¹ Two recent systematic reviews have demonstrated no difference in durability or functional outcome between APT and MBT components.^1,2 Despite this data, APT components continue to be used uncommonly in the United States. Improved technical ease and the theoretical advantages of modularity are likely responsible for the continued popularity of MBT implants despite the fact that APT implants cost considerably less than their MBT counterparts.

The importance of cost control in TKA is increasing secondary to changing economic realities of healthcare and increasing prevalence of joint replacement. Payers are seeking ways to ensure quality care at more affordable reimbursement rates. Surgeons play a critical role in cost containment and may soon be incentivized to make cost-effective decisions under proposed gainsharing programs. Implants account for a substantial portion of hospital costs for knee replacement and have been suggested as an essential part of cost control.³ As such, surgeons in the United States will probably need to factor in value when selecting implants and be required to justify the additional cost of “premium” implants.

Given recent systemic reviews concluding both equivalent effectiveness and survivorship, the APT component would appear to be inherently cost-effective when compared with an MBT design. However, the degree to which this implant is cost-effective has been difficult to quantify. The purpose of this study is to take a novel approach to examine the cost-effectiveness of APT components by determining what theoretical difference in revision rate would make modular MBT implants a more cost-effective intervention using our institutional cost data.

MATERIALS AND METHODS

A Markov decision model was used to evaluate the cost-effectiveness of APT components.⁴ A Markov decision model is a mathematical framework for modeling decision making in situations where outcomes are partly random and partly under the control of a decision maker. They are powerful tools for determining the best solution from all feasible solutions to a given problem. A decision model was constructed (Figure 1) to depict patients with arthritis of the knee being treated with either APT or MBT implants in a fashion similar to previously published models.⁵ At each point of a patient’s health status in the 20 years following surgery, they are either considered well after total knee replacement, well after revision surgery, or dead. Patients transition through the decision tree and pass through different states according to the probability of each event occurring, a process that is discussed further below. A utility value, measured in quality-adjusted life years (QALYs), and a cost are assigned to every health state and both primary and revision procedures within the model. The model is designed to determine the maximum failure rate for which the APT is the more cost-effective option.

The model probabilities used for survival and mortality following TKA were adapted from those published previously in the literature.⁵ A utility value was assigned to each health state. The utility after initial surgery was set to 0.83 and utility after revision was set to 0.6.⁵ These values were obtained from the Swedish Registry and Tufts Cost-Effectiveness Registry, respectively. We also included a disutility of -0.1 for the first year after surgery and -0.2 for the first year after revision, to account for the disutility of undergoing surgery and the post-surgery recovery. Disutilities represent the negative preference patients have for a particular health state or outcome, such as primary or revision knee arthroplasty.⁵ It is assumed that there is a higher morbidity associated with revision arthroplasty vs primary arthroplasty and, thus, has a higher disutility value assigned to it.

We assumed the age at the initial surgery to be 65 years. Age-specific mortality rates were taken from the 2007 United States Life Tables published by the Centers for Disease Control and Prevention.⁶ An additional probability of .007 of dying during the surgery or postoperative from the initial surgery and a probability of .011 from the revision was included.

Costs for the surgery were obtained from the University of Virginia’s billing department. We obtained the average cost for the diagnosis-related group in 2012. The cost of primary knee replacement was $17,578.06 with MBT implants. We subtracted institutional cost savings for the APT that could be achieved to obtain a cost of $16,272.10 for the APT. The cost of revision was $21,650.34 and assumed to be the same regardless of the type of initial surgery. A 3% discount rate was used.

The costs, QALYs, and probabilities were then used to compute cost-effectiveness ratios, or the cost per additional QALY, of the 2 options. Unlike previous models published in the orthopedic literature, we assumed a constant probability of revision for the MBT. We initially assumed a 1.0% probability of failure per year for the MBT implant. We then determined what revision rate for the APT would be necessary to be cost equivalent with the MBT. A sensitivity analysis was performed to examine the impact of varying assumptions regarding the rate of revision.

Continue to: Results...

 

RESULTS

Under our institutional cost data and model assumptions with a 1% annual failure rate for MBT implants, an annual failure rate of 1.6% for APT components would be required to achieve equivalency in cost. Over a 20-year period, a failure rate of >27% for the APT component would be necessary to achieve equivalent cost compared with the proposed failure rate of 18% with MBT components.

A two-way sensitivity analysis for probabilities of failure was performed to compare revision probabilities of the APT with those of MBT components. The preferred strategy graph is included in Figure 2. This graph shows how varying annual revision rates for both the APT and MBT would impact which option would be preferable. For example, on the graph, an annual failure rate of 1.6% for APT implants would be cost equivalent to a 0.1% annual failure rate for MBT implants at 20 years. A 2.0% annual failure rate for the APT would be equivalent to a 1.4% annual failure rate for the MBT, and a 2.5% failure rate for the APT would be equivalent to a 1.8% MBT failure rate. Holding the APT failure rate constant at 2.5%, any MBT failure rate <1.8% would make the MBT the more cost-effective option, whereas a failure rate >1.8% would make the MBT less cost-effective than the APT. For probability combinations that fall in the lower right area of Figure 2, the APT is preferable, and for probability combinations that fall in the upper left area, the MBT is preferable. The line separating the 2 areas is where 1 would be indifferent, such that the cost per additional QALY is the same for both procedures.

DISCUSSION

In light of the current economic climate and push for cost savings in the United States healthcare system, orthopedic surgeons must increasingly understand the realities of cost and the role it plays in the assessment of new technology. This concept is especially true of TKA as it becomes an increasingly common operative intervention. Utilizing cost savings techniques while ensuring quality outcomes is something that needs to be championed by healthcare providers.

Ideally, the introduction of a new medical technology that is more expensive than preexisting technology should lead to improved outcomes. Multiple randomized radiostereometric and clinical outcome studies looking at failure rates of APT compared with MBT have consistently suggested equivalence or superiority of the APT design when modern round-on-round implant designs are utilized.^7-17 Two recent systematic reviews demonstrated that APT components were equivalent to MBT components regarding both revision rates and clinical scores.^1,18 Given these results, it seems that the increased use of the APT design could save the healthcare system substantial amounts of money without compromising outcomes. For example, in 2006 Muller and colleagues^19. proposed a possible cost savings of approximately 39 million dollars per year across England and Wales, if just 50% of the 70,000 TKAs performed annually used APTs. Our study, which helps quantify the potential cost-effectiveness of the APT design in terms of revision rates, should help further support this debate and provide a framework for the evaluation of new technology.

It should be noted that the results of this current study are based on both assumptions and generalizations. Institutional cost data is known to vary widely among institutions and our conclusions regarding comparable revision rates would change with different cost inputs. We are also unable to take into account individual patients, surgeons, or specific implant factors. It is very difficult to place a price on quality-adjusted life years and negative repercussions with revision surgery. Furthermore, speaking specifically about surgical technique, each surgeon has his/her own preference when performing TKA. There is a lack of intraoperative flexibility when using monoblock tibial components that many surgeons may find undesirable. A surgeon is unable to adjust the thickness of the polyethylene insert after cementation of metal implants. Finally, we are aware that cost-effectiveness analyses cannot take the place of rational clinical decision making when evaluating an individual patient for TKA. Patient age, body mass index, and deformity are all factors that may dictate the use of MBTs in an attempt to improve outcomes.

The results of this analysis help quantify the cost-effectiveness of the APT. Given the additional cost, the MBT design would have to lower revision rates substantially when compared with the APT design to be considered cost-effective. Multiple clinical studies have not shown this to be the case. Further studies are required to help guide clinical decision making and define the role of APT components in TKA.

ABSTRACT

The importance of cost control in total knee arthroplasty is increasing in the United States secondary to both changing economic realities of healthcare and the increasing prevalence of joint replacement.

Surgeons play a critical role in cost containment and may soon be incentivized to make cost-effective decisions under proposed gainsharing programs. The purpose of this study is to examine the cost-effectiveness of all-polyethylene tibial (APT) components and determine what difference in revision rate would make modular metal-backed tibial (MBT) implants a more cost-effective intervention.

Markov models were constructed using variable implant failure rates and previously published probabilities. Cost data were obtained from both our institution and published United States implant list prices, and modeled with a 3.0% discount rate. The decision tree was continued over a 20-year timeframe.

Using our institutional cost data and model assumptions with a 1.0% annual failure rate for MBT components, an annual failure rate of 1.6% for APT components would be required to achieve equivalency in cost. Over a 20-year period, a failure rate of >27% for the APT component would be necessary to achieve equivalent cost compared with the proposed failure rate of 18% with MBT components. A sensitivity analysis was performed with different assumptions for MBT annual failure rates.

Given our assumptions, the APT component is cost-saving if the excess cumulative revision rate increases by <9% in 20 years compared with that of the MBT implant. Surgeons, payers, and hospitals should consider this approach when evaluating implants. Consideration should also be given to the decreased utility associated with revision surgery.

Continue to: All-polythylene tibial implants...

 

All-polyethylene tibial (APT) implants have been available for use in total knee arthroplasty (TKA) for decades. Except for one particular implant design, APT implants have shown equivalent functional outcome and survivorship to metal-backed tibial (MBT) components.¹ Two recent systematic reviews have demonstrated no difference in durability or functional outcome between APT and MBT components.^1,2 Despite this data, APT components continue to be used uncommonly in the United States. Improved technical ease and the theoretical advantages of modularity are likely responsible for the continued popularity of MBT implants despite the fact that APT implants cost considerably less than their MBT counterparts.

The importance of cost control in TKA is increasing secondary to changing economic realities of healthcare and increasing prevalence of joint replacement. Payers are seeking ways to ensure quality care at more affordable reimbursement rates. Surgeons play a critical role in cost containment and may soon be incentivized to make cost-effective decisions under proposed gainsharing programs. Implants account for a substantial portion of hospital costs for knee replacement and have been suggested as an essential part of cost control.³ As such, surgeons in the United States will probably need to factor in value when selecting implants and be required to justify the additional cost of “premium” implants.

Given recent systemic reviews concluding both equivalent effectiveness and survivorship, the APT component would appear to be inherently cost-effective when compared with an MBT design. However, the degree to which this implant is cost-effective has been difficult to quantify. The purpose of this study is to take a novel approach to examine the cost-effectiveness of APT components by determining what theoretical difference in revision rate would make modular MBT implants a more cost-effective intervention using our institutional cost data.

MATERIALS AND METHODS

A Markov decision model was used to evaluate the cost-effectiveness of APT components.⁴ A Markov decision model is a mathematical framework for modeling decision making in situations where outcomes are partly random and partly under the control of a decision maker. They are powerful tools for determining the best solution from all feasible solutions to a given problem. A decision model was constructed (Figure 1) to depict patients with arthritis of the knee being treated with either APT or MBT implants in a fashion similar to previously published models.⁵ At each point of a patient’s health status in the 20 years following surgery, they are either considered well after total knee replacement, well after revision surgery, or dead. Patients transition through the decision tree and pass through different states according to the probability of each event occurring, a process that is discussed further below. A utility value, measured in quality-adjusted life years (QALYs), and a cost are assigned to every health state and both primary and revision procedures within the model. The model is designed to determine the maximum failure rate for which the APT is the more cost-effective option.

The model probabilities used for survival and mortality following TKA were adapted from those published previously in the literature.⁵ A utility value was assigned to each health state. The utility after initial surgery was set to 0.83 and utility after revision was set to 0.6.⁵ These values were obtained from the Swedish Registry and Tufts Cost-Effectiveness Registry, respectively. We also included a disutility of -0.1 for the first year after surgery and -0.2 for the first year after revision, to account for the disutility of undergoing surgery and the post-surgery recovery. Disutilities represent the negative preference patients have for a particular health state or outcome, such as primary or revision knee arthroplasty.⁵ It is assumed that there is a higher morbidity associated with revision arthroplasty vs primary arthroplasty and, thus, has a higher disutility value assigned to it.

We assumed the age at the initial surgery to be 65 years. Age-specific mortality rates were taken from the 2007 United States Life Tables published by the Centers for Disease Control and Prevention.⁶ An additional probability of .007 of dying during the surgery or postoperative from the initial surgery and a probability of .011 from the revision was included.

Costs for the surgery were obtained from the University of Virginia’s billing department. We obtained the average cost for the diagnosis-related group in 2012. The cost of primary knee replacement was $17,578.06 with MBT implants. We subtracted institutional cost savings for the APT that could be achieved to obtain a cost of $16,272.10 for the APT. The cost of revision was $21,650.34 and assumed to be the same regardless of the type of initial surgery. A 3% discount rate was used.

The costs, QALYs, and probabilities were then used to compute cost-effectiveness ratios, or the cost per additional QALY, of the 2 options. Unlike previous models published in the orthopedic literature, we assumed a constant probability of revision for the MBT. We initially assumed a 1.0% probability of failure per year for the MBT implant. We then determined what revision rate for the APT would be necessary to be cost equivalent with the MBT. A sensitivity analysis was performed to examine the impact of varying assumptions regarding the rate of revision.

Continue to: Results...

 

RESULTS

Under our institutional cost data and model assumptions with a 1% annual failure rate for MBT implants, an annual failure rate of 1.6% for APT components would be required to achieve equivalency in cost. Over a 20-year period, a failure rate of >27% for the APT component would be necessary to achieve equivalent cost compared with the proposed failure rate of 18% with MBT components.

A two-way sensitivity analysis for probabilities of failure was performed to compare revision probabilities of the APT with those of MBT components. The preferred strategy graph is included in Figure 2. This graph shows how varying annual revision rates for both the APT and MBT would impact which option would be preferable. For example, on the graph, an annual failure rate of 1.6% for APT implants would be cost equivalent to a 0.1% annual failure rate for MBT implants at 20 years. A 2.0% annual failure rate for the APT would be equivalent to a 1.4% annual failure rate for the MBT, and a 2.5% failure rate for the APT would be equivalent to a 1.8% MBT failure rate. Holding the APT failure rate constant at 2.5%, any MBT failure rate <1.8% would make the MBT the more cost-effective option, whereas a failure rate >1.8% would make the MBT less cost-effective than the APT. For probability combinations that fall in the lower right area of Figure 2, the APT is preferable, and for probability combinations that fall in the upper left area, the MBT is preferable. The line separating the 2 areas is where 1 would be indifferent, such that the cost per additional QALY is the same for both procedures.

DISCUSSION

In light of the current economic climate and push for cost savings in the United States healthcare system, orthopedic surgeons must increasingly understand the realities of cost and the role it plays in the assessment of new technology. This concept is especially true of TKA as it becomes an increasingly common operative intervention. Utilizing cost savings techniques while ensuring quality outcomes is something that needs to be championed by healthcare providers.

Ideally, the introduction of a new medical technology that is more expensive than preexisting technology should lead to improved outcomes. Multiple randomized radiostereometric and clinical outcome studies looking at failure rates of APT compared with MBT have consistently suggested equivalence or superiority of the APT design when modern round-on-round implant designs are utilized.^7-17 Two recent systematic reviews demonstrated that APT components were equivalent to MBT components regarding both revision rates and clinical scores.^1,18 Given these results, it seems that the increased use of the APT design could save the healthcare system substantial amounts of money without compromising outcomes. For example, in 2006 Muller and colleagues^19. proposed a possible cost savings of approximately 39 million dollars per year across England and Wales, if just 50% of the 70,000 TKAs performed annually used APTs. Our study, which helps quantify the potential cost-effectiveness of the APT design in terms of revision rates, should help further support this debate and provide a framework for the evaluation of new technology.

It should be noted that the results of this current study are based on both assumptions and generalizations. Institutional cost data is known to vary widely among institutions and our conclusions regarding comparable revision rates would change with different cost inputs. We are also unable to take into account individual patients, surgeons, or specific implant factors. It is very difficult to place a price on quality-adjusted life years and negative repercussions with revision surgery. Furthermore, speaking specifically about surgical technique, each surgeon has his/her own preference when performing TKA. There is a lack of intraoperative flexibility when using monoblock tibial components that many surgeons may find undesirable. A surgeon is unable to adjust the thickness of the polyethylene insert after cementation of metal implants. Finally, we are aware that cost-effectiveness analyses cannot take the place of rational clinical decision making when evaluating an individual patient for TKA. Patient age, body mass index, and deformity are all factors that may dictate the use of MBTs in an attempt to improve outcomes.

The results of this analysis help quantify the cost-effectiveness of the APT. Given the additional cost, the MBT design would have to lower revision rates substantially when compared with the APT design to be considered cost-effective. Multiple clinical studies have not shown this to be the case. Further studies are required to help guide clinical decision making and define the role of APT components in TKA.

ABSTRACT

The importance of cost control in total knee arthroplasty is increasing in the United States secondary to both changing economic realities of healthcare and the increasing prevalence of joint replacement.

Surgeons play a critical role in cost containment and may soon be incentivized to make cost-effective decisions under proposed gainsharing programs. The purpose of this study is to examine the cost-effectiveness of all-polyethylene tibial (APT) components and determine what difference in revision rate would make modular metal-backed tibial (MBT) implants a more cost-effective intervention.

Markov models were constructed using variable implant failure rates and previously published probabilities. Cost data were obtained from both our institution and published United States implant list prices, and modeled with a 3.0% discount rate. The decision tree was continued over a 20-year timeframe.

Using our institutional cost data and model assumptions with a 1.0% annual failure rate for MBT components, an annual failure rate of 1.6% for APT components would be required to achieve equivalency in cost. Over a 20-year period, a failure rate of >27% for the APT component would be necessary to achieve equivalent cost compared with the proposed failure rate of 18% with MBT components. A sensitivity analysis was performed with different assumptions for MBT annual failure rates.

Given our assumptions, the APT component is cost-saving if the excess cumulative revision rate increases by <9% in 20 years compared with that of the MBT implant. Surgeons, payers, and hospitals should consider this approach when evaluating implants. Consideration should also be given to the decreased utility associated with revision surgery.

Continue to: All-polythylene tibial implants...

 

All-polyethylene tibial (APT) implants have been available for use in total knee arthroplasty (TKA) for decades. Except for one particular implant design, APT implants have shown equivalent functional outcome and survivorship to metal-backed tibial (MBT) components.¹ Two recent systematic reviews have demonstrated no difference in durability or functional outcome between APT and MBT components.^1,2 Despite this data, APT components continue to be used uncommonly in the United States. Improved technical ease and the theoretical advantages of modularity are likely responsible for the continued popularity of MBT implants despite the fact that APT implants cost considerably less than their MBT counterparts.

The importance of cost control in TKA is increasing secondary to changing economic realities of healthcare and increasing prevalence of joint replacement. Payers are seeking ways to ensure quality care at more affordable reimbursement rates. Surgeons play a critical role in cost containment and may soon be incentivized to make cost-effective decisions under proposed gainsharing programs. Implants account for a substantial portion of hospital costs for knee replacement and have been suggested as an essential part of cost control.³ As such, surgeons in the United States will probably need to factor in value when selecting implants and be required to justify the additional cost of “premium” implants.

Given recent systemic reviews concluding both equivalent effectiveness and survivorship, the APT component would appear to be inherently cost-effective when compared with an MBT design. However, the degree to which this implant is cost-effective has been difficult to quantify. The purpose of this study is to take a novel approach to examine the cost-effectiveness of APT components by determining what theoretical difference in revision rate would make modular MBT implants a more cost-effective intervention using our institutional cost data.

MATERIALS AND METHODS

A Markov decision model was used to evaluate the cost-effectiveness of APT components.⁴ A Markov decision model is a mathematical framework for modeling decision making in situations where outcomes are partly random and partly under the control of a decision maker. They are powerful tools for determining the best solution from all feasible solutions to a given problem. A decision model was constructed (Figure 1) to depict patients with arthritis of the knee being treated with either APT or MBT implants in a fashion similar to previously published models.⁵ At each point of a patient’s health status in the 20 years following surgery, they are either considered well after total knee replacement, well after revision surgery, or dead. Patients transition through the decision tree and pass through different states according to the probability of each event occurring, a process that is discussed further below. A utility value, measured in quality-adjusted life years (QALYs), and a cost are assigned to every health state and both primary and revision procedures within the model. The model is designed to determine the maximum failure rate for which the APT is the more cost-effective option.

The model probabilities used for survival and mortality following TKA were adapted from those published previously in the literature.⁵ A utility value was assigned to each health state. The utility after initial surgery was set to 0.83 and utility after revision was set to 0.6.⁵ These values were obtained from the Swedish Registry and Tufts Cost-Effectiveness Registry, respectively. We also included a disutility of -0.1 for the first year after surgery and -0.2 for the first year after revision, to account for the disutility of undergoing surgery and the post-surgery recovery. Disutilities represent the negative preference patients have for a particular health state or outcome, such as primary or revision knee arthroplasty.⁵ It is assumed that there is a higher morbidity associated with revision arthroplasty vs primary arthroplasty and, thus, has a higher disutility value assigned to it.

We assumed the age at the initial surgery to be 65 years. Age-specific mortality rates were taken from the 2007 United States Life Tables published by the Centers for Disease Control and Prevention.⁶ An additional probability of .007 of dying during the surgery or postoperative from the initial surgery and a probability of .011 from the revision was included.

Costs for the surgery were obtained from the University of Virginia’s billing department. We obtained the average cost for the diagnosis-related group in 2012. The cost of primary knee replacement was $17,578.06 with MBT implants. We subtracted institutional cost savings for the APT that could be achieved to obtain a cost of $16,272.10 for the APT. The cost of revision was $21,650.34 and assumed to be the same regardless of the type of initial surgery. A 3% discount rate was used.

The costs, QALYs, and probabilities were then used to compute cost-effectiveness ratios, or the cost per additional QALY, of the 2 options. Unlike previous models published in the orthopedic literature, we assumed a constant probability of revision for the MBT. We initially assumed a 1.0% probability of failure per year for the MBT implant. We then determined what revision rate for the APT would be necessary to be cost equivalent with the MBT. A sensitivity analysis was performed to examine the impact of varying assumptions regarding the rate of revision.

Continue to: Results...

 

RESULTS

Under our institutional cost data and model assumptions with a 1% annual failure rate for MBT implants, an annual failure rate of 1.6% for APT components would be required to achieve equivalency in cost. Over a 20-year period, a failure rate of >27% for the APT component would be necessary to achieve equivalent cost compared with the proposed failure rate of 18% with MBT components.

A two-way sensitivity analysis for probabilities of failure was performed to compare revision probabilities of the APT with those of MBT components. The preferred strategy graph is included in Figure 2. This graph shows how varying annual revision rates for both the APT and MBT would impact which option would be preferable. For example, on the graph, an annual failure rate of 1.6% for APT implants would be cost equivalent to a 0.1% annual failure rate for MBT implants at 20 years. A 2.0% annual failure rate for the APT would be equivalent to a 1.4% annual failure rate for the MBT, and a 2.5% failure rate for the APT would be equivalent to a 1.8% MBT failure rate. Holding the APT failure rate constant at 2.5%, any MBT failure rate <1.8% would make the MBT the more cost-effective option, whereas a failure rate >1.8% would make the MBT less cost-effective than the APT. For probability combinations that fall in the lower right area of Figure 2, the APT is preferable, and for probability combinations that fall in the upper left area, the MBT is preferable. The line separating the 2 areas is where 1 would be indifferent, such that the cost per additional QALY is the same for both procedures.

DISCUSSION

In light of the current economic climate and push for cost savings in the United States healthcare system, orthopedic surgeons must increasingly understand the realities of cost and the role it plays in the assessment of new technology. This concept is especially true of TKA as it becomes an increasingly common operative intervention. Utilizing cost savings techniques while ensuring quality outcomes is something that needs to be championed by healthcare providers.

Ideally, the introduction of a new medical technology that is more expensive than preexisting technology should lead to improved outcomes. Multiple randomized radiostereometric and clinical outcome studies looking at failure rates of APT compared with MBT have consistently suggested equivalence or superiority of the APT design when modern round-on-round implant designs are utilized.^7-17 Two recent systematic reviews demonstrated that APT components were equivalent to MBT components regarding both revision rates and clinical scores.^1,18 Given these results, it seems that the increased use of the APT design could save the healthcare system substantial amounts of money without compromising outcomes. For example, in 2006 Muller and colleagues^19. proposed a possible cost savings of approximately 39 million dollars per year across England and Wales, if just 50% of the 70,000 TKAs performed annually used APTs. Our study, which helps quantify the potential cost-effectiveness of the APT design in terms of revision rates, should help further support this debate and provide a framework for the evaluation of new technology.

It should be noted that the results of this current study are based on both assumptions and generalizations. Institutional cost data is known to vary widely among institutions and our conclusions regarding comparable revision rates would change with different cost inputs. We are also unable to take into account individual patients, surgeons, or specific implant factors. It is very difficult to place a price on quality-adjusted life years and negative repercussions with revision surgery. Furthermore, speaking specifically about surgical technique, each surgeon has his/her own preference when performing TKA. There is a lack of intraoperative flexibility when using monoblock tibial components that many surgeons may find undesirable. A surgeon is unable to adjust the thickness of the polyethylene insert after cementation of metal implants. Finally, we are aware that cost-effectiveness analyses cannot take the place of rational clinical decision making when evaluating an individual patient for TKA. Patient age, body mass index, and deformity are all factors that may dictate the use of MBTs in an attempt to improve outcomes.

The results of this analysis help quantify the cost-effectiveness of the APT. Given the additional cost, the MBT design would have to lower revision rates substantially when compared with the APT design to be considered cost-effective. Multiple clinical studies have not shown this to be the case. Further studies are required to help guide clinical decision making and define the role of APT components in TKA.

ABSTRACT

The purpose of this study is to describe the rate of return to the operating room (OR) following microfracture (MFX), autologous chondrocyte implantation (ACI), osteochondral autograft transplantation (OATS), and osteochondral allograft (OCA) procedures at 90 days, 1 year, and 2 years. Current Procedural Terminology codes for all patients undergoing MFX, ACI, OATS, and OCA were used to search a prospectively collected, commercially available private payer insurance company database from 2007 to 2011. Within 90 days, 1 year, and 2 years after surgery, the database was searched for the occurrence of these same patients undergoing knee diagnostic arthroscopy with biopsy, lysis of adhesions, synovectomy, arthroscopy for infection or lavage, arthroscopy for removal of loose bodies, chondroplasty, MFX, ACI, OATS, OCA, and/or knee arthroplasty. Descriptive statistical analysis and contingency table analysis were performed. A total of 47,207 cartilage procedures were performed from 2007 to 2011, including 43,576 MFX, 640 ACI, 386 open OATS, 997 arthroscopic OATS, 714 open OCA, and 894 arthroscopic OCA procedures. The weighted average reoperation rates for all procedures were 5.87% at 90 days, 11.94% at 1 year, and 14.90% at 2 years following the index cartilage surgery. At 2 years, patients who underwent MFX, ACI, OATS, OCA had reoperation rates of 14.65%, 29.69%, 8.82%, and 12.22%, respectively. There was a statistically significantly increased risk for ACI return to OR within all intervals (P < .0001); however, MFX had a greater risk factor (P < .0001) for conversion to arthroplasty. There was no difference in failure/revision rates between the restorative treatment options. With a large US commercial insurance database from 2007 to 2011, reparative procedures were favored for chondral injuries, but yielded an increased risk for conversion to arthroplasty. There was no difference in failure/revision rates between the restorative approaches, yet cell-based approaches yielded a significantly increased risk for a return to the OR.

Continue to: Symptomatic, full-thickness articular cartilage

 

Symptomatic, full-thickness articular cartilage defects in the knee are difficult to manage, particularly in the young, athletic patient population. Fortunately, a variety of cartilage repair (direct repair of the cartilage or those procedures which attempt to generate fibrocartilage) and restoration (those aimed at restoring hyaline cartilage) procedures are available, with encouraging short- and long-term clinical outcomes. After failure of nonoperative management, several surgical options are available for treating symptomatic focal chondral defects, including microfracture (MFX), autologous chondrocyte implantation (ACI), osteochondral autograft transplantation (OATS), and open and arthroscopic osteochondral allograft (OCA) transplantation procedures.^1,2 When appropriately indicated, each of these techniques has demonstrated good to excellent clinical outcomes with respect to reducing pain and improving function.^3-5

While major complications following cartilage surgery are uncommon, the need for reoperation following an index articular cartilage operation is poorly understood. Recently, McCormick and colleagues6 found that reoperation within the first 2 years following meniscus allograft transplantation (MAT) is associated with an increased likelihood of revision MAT or future arthroplasty. Given the association between early reoperation following meniscus restoration surgery and subsequent failure, an improved understanding of the epidemiology and implications of reoperations following cartilage restoration surgery is warranted. Further, in deciding which treatment option is best suited to a particular patient, the rate of return to the operating room (OR) should be taken into consideration, as this could potentially influence surgical decision-making as to which procedure to perform, especially in value-based care decision-making environments.

The purpose of this study is to describe the rate of return to the OR for knee procedures following cartilage restoration at intervals of 90 days, 1 year, and 2 years across a large-scale US patient database. The authors hypothesize that the rate of return to the OR following knee cartilage repair or restoration procedures will be under 20% during the first post-operative year, with increasing reoperation rates over time. A secondary hypothesis is that there will be no difference in reoperation rates according to sex, but that younger patients (those younger than 40 years) will have higher reoperation rates than older patients.

METHODS

We performed a retrospective analysis of a prospectively collected, large-scale, and commercially available private payer insurance company database (PearlDiver) from 2007 to 2011. The PearlDiver database is a Health Insurance Portability and Accountability Act (HIPAA) compliant, publicly available national database consisting of a collection of private payer records, with United Health Group representing the contributing health plan. The database has more than 30 million patient records and contains Current Procedural Terminology (CPT) and International Classification of Diseases, Ninth Revision (ICD-9) codes related to orthopedic procedures. From 2007 to 2011, the private payer database captured between 5.9 million and 6.2 million patients per year.

Our search was based on the CPT codes for MFX (29879), ACI (27412), OATS (29866, 29867), and OCA (27415, 27416). Return to the OR for revision surgery for the above-mentioned procedures was classified as patients with a diagnosis of diagnostic arthroscopy with biopsy (CPT 29870), lysis of adhesions (CPT 29884), synovectomy (29875, 29876), arthroscopy for infection or lavage (CPT 29871), arthroscopy for removal of loose bodies (29874), chondroplasty (29877), unicompartmental knee arthroplasty (27446), total knee arthroplasty (27447), and/or patellar arthroplasty (27438). Patient records were followed for reoperations occurring within 90 days, 1 year, and 2 years after the index cartilage procedure. All data were compared based on patient age and sex.

Table 1. Breakdown of MFX, ACI, OATS, and OCA Procedures by Sex
	MFX	ACI	Open OATS	Arthroscopic OATS	Open OCA	Arthroscopic OCA
Females	20,589	276	167	401	275	350
Males	22,987	364	219	596	439	544
Total	43,576	640	386	997	714	894

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation.

Continue to: Statistical analysis...

 

STATISTICAL ANALYSIS

Statistical analysis of this study was primarily descriptive to demonstrate the incidence for each code at each time interval. One-way analysis of variance, Chi-square analysis, and contingency tables were used to compare the incidence of each type of procedure throughout the various time intervals. A P-value of < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS v.20 (International Business Machines).

RESULTS

A total of 47,207 cartilage procedures were performed from 2007 to 2011, including 43,576 MFX (92.3%) 640 ACI (1.4%), 386 open OATS (0.82%), 997 arthroscopic OATS (2.11%), 714 open OCA (1.51%), and 894 arthroscopic OCA (1.89%) procedures. A summary of the procedures performed, broken down by age and sex, is provided in Tables 1 and 2. A total of 25,149 male patients (53.3%) underwent surgical procedures compared to 22,058 female patients (46.7%). For each category of procedure (MFX, ACI, OATS, OCA), there was a significantly higher proportion of males than females undergoing surgery (P < .0001 for all). Surgical treatment with MFX was consistently the most frequently performed surgery across all age groups (92.31%), while cell-based therapy with ACI was the least frequently performed procedure across all age ranges (1.36%). Restorative OATS and OCA techniques were performed with the greatest frequency in the 15-year-old to 19-year-old age group, but were not utilized in patients over 64 years of age (Table 2).

Table 2. Breakdown of MFX, ACI, OATS, and OCA Procedures by Age
Age (y)	MFX	ACI	OATS	OCA
10 to 14	572	22	74	47
15 to 19	1984	83	254	235
20 to 24	1468	54	140	144
25 to 29	1787	74	152	176
30 to 34	2824	114	152	204
35 to 39	4237	96	153	210
40 to 44	5441	103	166	217
45 to 49	7126	57	149	180
50 to 54	7004	25	83	140
55 to 59	6410	12	40	40
60 to 64	4409	0	20	15
65 to 69	269	0	0	0
70 to 74	45	0	0	0
Total	43,576	640	1383	1608

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation.

A summary of all reoperation data is provided in Tables 3 to 7 and Figures 1 and 2. The weighted average reoperation rates for all procedures were 5.87% at 90 days, 11.94% at 1 year, and 14.90% at 2 years following the index cartilage surgery. Patients who underwent MFX had reoperation rates of 6.05% at 90 days, 11.80% at 1 year, and 14.65% at 2 years. Patients who underwent ACI had reoperation rates of 4.53% at 90 days, 23.28% at 1 year, and 29.69% at 2 years. Patients who had open and arthroscopic OATS had reoperation rates of 3.122% and 5.12% at 90 days, 6.74% and 8.53% at 1 year, and 7.51% and 10.13% at 2 years, respectively. Patients who underwent open and arthroscopic OCA had reoperation rates of 2.52% and 3.91% at 90 days, 7.14% and 6.60% at 1 year, and 13.59% and 10.85% at 2 years (Table 3). There was a statistically significantly increased risk for reoperation following ACI within all intervals compared to all other surgical techniques (P < .0001); however, MFX had a greater risk factor (P < .0001) for conversion to arthroplasty at 6.70%. There was no significant difference between failure rates (revision OATS/OCA or conversion to arthroplasty) between the restorative treatment options, with 14 failures for OATS (9.52% of reoperations at 2 years) compared to 22 failures for OCA (12.7% of reoperations at 2 years, P = .358). Among the entire cohort of cartilage surgery patients, arthroscopic chondroplasty was the most frequent procedure performed at the time of reoperation at all time points assessed, notably accounting for 33.08% of reoperations 2 years following microfracture, 51.58% of reoperations at 2 years following ACI, 53.06% of reoperations at 2 years following OATS, and 54.07% of reoperations at 2 years following OCA (Figure 3, Tables 4–7).

Table 3. Comparison of Return to OR Following MFX, ACI, OCA, and OATS
Procedure	Total No. of Cases in Study Period	No. of Reoperations at 90 Days	Return to OR Rate at 90 Days	No. of Reoperations at 1 Year	Return to OR Rate at 1 Year	No. of Reoperations at 2 Years	Return to OR Rate at 2 Years
MFX	43,576	2636	6.05%	5142	11.80%	6385	14.65%
ACI	640	29	4.53%	149	23.28%	190	29.69%
Open OATS	386	12	3.12%	26	6.74%	29	7.51%
Arthroscopic OATS	997	51	5.12%	85	8.53%	101	10.13%
Open OCA	714	18	2.52%	51	7.14%	97	13.59%
Arthroscopic OCA	894	161	3.91%	59	6.60%	97	10.85%
Weighted average for all procedures			5.87%		11.94%		14.90%

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation; OR, operating room.

Table 4. Rate of Return to OR Following MFX (n = 43,574)
Procedure	CPT Code	90 Days	1 Year	2 Years
Knee arthroscopy	29870	54	122	162
Knee arthroscopic drainage and lavage	29871	84	102	104
Arthroscopic adhesions débridement	29874	300	468	549
Arthroscopic synovectomy	29875	324	528	611
Major arthroscopic synovectomy	29876	557	926	1087
Knee arthroscopic chondroplasty	29877	1063	1722	2112
Arthroscopic lysis of adhesions	29884	61	129	171
Patellar arthroplasty	27438	0	38	49
Medial or lateral knee arthroplasty	27446	51	242	328
Medial and lateral knee arthroplasty	27447	142	865	1212
Total		2636	5142	6385
Return to OR		6.05%	11.80%	14.65%

Abbreviations: CPT, Current Procedural Terminology; MFX, microfracture; OR, operating room.

Table 5. Rate of Return to OR Following ACI (n = 640)
Procedure	CPT Code	90 Daysa	1 Yeara	2 Yearsa
Revision ACI	27412	29	33	35
Knee arthroscopy	29870	-1	-1	-1
Knee arthroscopic drainage and lavage	29871	-1	-1	-1
Arthroscopic adhesions débridement	29874	0	-1	-1
Arthroscopic synovectomy	29875	-1	-1	-1
Major arthroscopic synovectomy	29876	-1	12	20
Knee arthroscopic chondroplasty	29877	-1	71	98
Arthroscopic lysis of adhesions	29884	-1	33	37
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	-1	-1
Medial and lateral knee arthroplasty	27447	0	-1	-1
Total		29	149	190
Return to OR		4.53%	23.28%	29.69%

^aA -1 denotes No. <11 within the PearlDiver database, and exact numbers are not reported due to patient privacy considerations.

Abbreviations: ACI, autologous chondrocyte implantation; CPT, Current Procedural Terminology; OR, operating room.

Table 6. Rate of Return to OR Following OATS (n = 1320)
Procedure	CPT Code	90 Days	1 Year	2 Years
Knee arthroscopy	29870	0	0	0
Knee arthroscopic drainage and lavage	29871	0	0	0
Arthroscopic adhesions débridement	29874	0	12	13
Arthroscopic synovectomy	29875	0	0	14
Major arthroscopic synovectomy	29876	16	25	28
Knee arthroscopic chondroplasty	29877	17	58	78
Arthroscopic lysis of adhesions	29884	0	0	0
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	0	0
Medial and lateral knee arthroplasty	27447	0	0	14
Total		33	95	147
Return to OR		2.50%	7.20%	11.14%

Abbreviations: CPT, Current Procedural Terminology; OATS, osteochondral autograft transplantation; OR, operating room.

Table 7. Rate of Return to OR Following OCA Transplantation (n = 1531)
Procedure	CPT Code	90 Days	1 Year	2 Year
Knee arthroscopy	29870	0	0	0
Knee arthroscopic drainage and lavage	29871	0	0	0
Arthroscopic adhesions débridement	29874	0	15	19
Arthroscopic synovectomy	29875	0	0	0
Major arthroscopic synovectomy	29876	0	20	38
Knee arthroscopic chondroplasty	29877	22	59	93
Arthroscopic lysis of adhesions	29884	0	0	0
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	0	0
Medial and lateral knee arthroplasty	27447	0	0	22
Total		22	94	172
Return to OR		1.44%	6.14%	11.23%

Abbreviations: CPT, Current Procedural Terminology; OCA, osteochondral allograft; OR, operating room.

Continue to: Discussion...

 

 

DISCUSSION

The principle findings of this study demonstrate that there is an overall reoperation rate of 14.90% at 2 years following cartilage repair/restoration surgery, with the highest reoperation rates following MFX at 90 days, and ACI at both 1 year and 2 years following the index procedure. Also, patients undergoing index MFX as the index procedure have the highest risk for conversion to arthroplasty, reoperation rates for all cartilage surgeries increase over time, and arthroscopic chondroplasty is the most frequent procedure performed at the time of reoperation.

The management of symptomatic articular cartilage knee pathology is extremely challenging. With improvements in surgical technique, instrumentation, and clinical decision-making, indications are constantly evolving. Techniques that may work for “small” defects, though there is some debate as to what constitutes a “small” defect, are not necessarily going to be successful for larger defects, and this certainly varies depending on where the defect is located within the knee joint (distal femur vs patella vs trochlea, etc.). Recently, in a 2015 analysis of 3 level I or II studies, Miller and colleagues⁷ demonstrated both MFX and OATS to be viable, cost-effective, first-line treatment options for articular cartilage injuries, with similar clinical outcomes at 8.7 years. The authors noted cumulative reoperation rates of 29% among patients undergoing MFX compared to 13% among patients undergoing OATS. While ACI and OCA procedures were not included in their study, the reported reoperation rates of 29% following MFX and 13% following OATS at nearly 10 years suggest a possible increased need for reoperation following MFX over time (approximately 15% at 2 years in our study) and a stable rate of reoperation following OATS (approximately 11% at 2 years in our study). This finding is significant, as one of the goals with these procedures is to deliver effective, long-lasting pain relief and restoration of function. Interestingly, in this study, restorative OATS and OCA techniques were performed with the greatest frequency in the 15-year-old to 19-year-old age group, but were not performed in patients older than 64 years. This may be explained by the higher prevalence of acute traumatic injuries and osteochondritis dissecans diagnoses in younger patients compared with older patients, as these diagnoses are more often indicated to undergo restorative procedures as opposed to marrow stimulation.

In a 2016 systematic review of 20 studies incorporating 1117 patients, Campbell and colleagues⁸ assessed return-to-play rates following MFX, ACI, OATS, and OCA. The authors noted that return to sport (RTS) rates were greatest following OATS (89%), followed by OCA (88%), ACI (84%), and MFX (75%). Positive prognostic factors for RTS included younger age, shorter duration of preoperative symptoms, no history of prior ipsilateral knee surgery, and smaller chondral defects. Reoperation rates between the 4 techniques were not statistically compared in their study. Interestingly, in 2013, Chalmers and colleagues⁹ conducted a separate systematic review of 20 studies comprising 1375 patients undergoing MFX, ACI, or OATS. In their study, the authors found significant advantages following ACI and OATS compared to MFX with respect to patient-reported outcome scores but noted significantly faster RTS rates with MFX. Reoperation rates were noted to be similar between the 3 procedures (25% for ACI, 21% for MFX, and 28% for OATS) at an average 3.7 years following the index procedure. When considering these 2 systematic reviews together, despite a faster RTS rate following MFX, a greater proportion of patients seem to be able to RTS over time following other procedures such as OATS, OCA, and ACI. Unfortunately, these reviews do not provide insight as to the role, if any, of reoperation on return to play rates nor on overall clinical outcome scores on patients undergoing articular cartilage surgery. However, this information is valuable when counseling athletes who are in season and would like to RTS as soon as possible as opposed to those who do not have tight time constraints for when they need to RTS.

Regardless of the cartilage technique chosen, the goals of surgery remain similar—to reduce pain and improve function. For athletes, the ultimate goal is to return to the same level of play that the athlete was able to achieve prior to injury. Certainly, the need for reoperation following a cartilage surgery has implications on pain, function, and ability to RTS. Our review of nearly 50,000 cartilage surgeries demonstrates that reoperations following cartilage repair surgery are not uncommon, with a rate of 14.90% at 2 years, and that while reoperation rates are the highest following ACI, the rate of conversion to knee arthroplasty is highest following MFX. Due to the limitations of the PearlDiver database, it is not possible to determine the clinical outcomes of patients undergoing reoperation following cartilage surgery, but certainly, given these data, reoperation is clearly not necessarily indicative of clinical failure. This is highlighted by the fact that the most common procedure performed at the time of reoperation is arthroscopic chondroplasty, which, despite being an additional surgical procedure, may be acceptable for patients who wish to RTS, particularly in the setting of an index ACI in which there may be graft hypertrophy. Ideally, additional studies incorporating a cost-effectiveness analysis of each of the procedures, incorporating reoperation rates as well as patient-reported clinical outcomes, would be helpful to truly determine the patient and societal implications of reoperation following cartilage repair/restoration.

Many of the advantages and disadvantages of the described cartilage repair/restoration procedures have been well described.^10-17 Microfracture is the most commonly utilized first-line repair/restoration option for small articular cartilage lesions, mainly due to its low cost, low morbidity, and relatively low level of difficulty.18 Despite these advantages, MFX is not without limitations, and the need for revision cartilage restoration and/or conversion to arthroplasty is concerning. In 2013, Salzmann and colleagues19 evaluated a cohort of 454 patients undergoing MFX for a symptomatic knee defect and noted a reoperation rate of 26.9% (n = 123) within 2 years of the index surgery, with risk factors for reoperation noted to include an increased number of pre-MFX ipsilateral knee surgeries, patellofemoral lesions, smoking, and lower preoperative numeric analog scale scores. The definition of reoperation in their study is unfortunately not described, and thus the extent of reoperation (arthroscopy to arthroplasty) is unclear. In a 2009 systematic review of 3122 patients (28 studies) undergoing MFX conducted by Mithoefer and colleagues,²⁰ revision rates were noted to range from 2% to 31% depending on the study analyzed, with increasing revision rates after 2 years. Unfortunately, the heterogeneity of the included studies makes it difficult to determine which patients tend to fail over time.

Continue to: OATS...

 

OATS is a promising cartilage restoration technique indicated for treatment of patients with large, uncontained chondral lesions, and/or lesions with both bone and cartilage loss.¹ OCA is similar to OATS but uses allograft tissue instead of autograft tissue and is typically considered a viable treatment option in larger lesions (>2 cm²).²¹ Cell-based ACI therapy has evolved substantially over the past decade and is now available as a third-generation model utilizing biodegradable 3-dimensional scaffolds seeded with chondrocytes. Reoperation rates following ACI can often be higher than those following other cartilage treatments, particularly given the known complication of graft hypertrophy and/or delamination. Harris and colleagues²² conducted a systematic review of 5276 subjects undergoing ACI (all generations), noting an overall reoperation rate of 33%, but a failure rate of 5.8% at an average of 22 months following ACI. Risk factors for reoperation included periosteal-based ACI as well as open (vs arthroscopic) ACI. In this study, we found a modestly lower return to OR rate of 29.69% at 2 years.

When the outcomes of patients undergoing OATS or OCA are compared to those of patients undergoing MFX or ACI, it can be difficult to interpret the results, as the indications for performing these procedures tend to be very different. Further, the reasons for reoperation, as well as the procedures performed at the time of reoperation, are often poorly described, making it difficult to truly quantify the risk of reoperation and the implications of reoperation for patients undergoing any of these index cartilage procedures.

Overall, in this database, the return to the OR rate approaches 15% at 2 years following cartilage surgery, with cell-based therapy demonstrating higher reoperation rates at 2 years, without the risk of conversion to arthroplasty. Reoperation rates appear to stabilize at 1 year following surgery and consist mostly of minor arthroscopic procedures. These findings can help surgeons counsel patients as to the rate and type of reoperations that can be expected following cartilage surgery. Additional research incorporating patient-reported outcomes and patient-specific risk factors are needed to complement these data as to the impact of reoperations on overall clinical outcomes. Further, studies incorporating 90-day, 1-year, and 2-year costs associated with cartilage surgery will help to determine which index procedure is the most cost effective over the short- and long-term.

LIMITATIONS

This study is not without limitations. The PearlDiver database is reliant upon accurate CPT and ICD-9 coding, which creates a potential for a reporting bias. The overall reliability of the analyses is dependent on the quality of the available data, which, as noted in previous PearlDiver studies,^18,23-28 may include inaccurate billing codes, miscoding, and/or non-coding by physicians as potential sources of error. At the time of this study, the PearlDiver database did not provide consistent data points on laterality, and thus it is possible that the reported rates of reoperation overestimate the true reoperation rate following a given procedure. Fortunately, the reoperation rates for each procedure analyzed in this database study are consistent with those previously presented in the literature. In addition, it is not uncommon for patients receiving one of these procedures to have previously been treated with one of the others. Due to the inherent limitations of the PearlDiver database, this study did not investigate concomitant procedures performed along with the index procedure, nor did it investigate confounding factors such as comorbidities. The PearlDiver database does not provide data on defect size, location within the knee, concomitant pathologies (eg, meniscus tear), prior surgeries, or patient comorbidities, and while important, these factors cannot be accounted for in our analysis. The inability to account for these important factors, particularly concomitant diagnoses, procedures, and lesion size/location, represents an important limitation of this study, as this is a source of selection bias and may influence the need for reoperation in a given patient. Despite these limitations, the results of this study are supported by previous and current literature. In addition, the PearlDiver database, as a HIPAA-compliant database, does not report exact numbers when the value of the outcome of interest is between 0 and 10, which prohibits analysis of any cartilage procedure performed in a cohort of patients greater than 1 and less than 11. Finally, while not necessarily a limitation, it should be noted that CPT 29879 is not specific for microfracture, as the code also includes abrasion arthroplasty and drilling. Due to the limitations of the methodology of searching the database for this code, it is unclear as to how many patients underwent actual microfracture vs abrasion arthroplasty.

CONCLUSION

Within a large US commercial insurance database from 2007 to 2011, reparative procedures were favored for chondral injuries, but yielded an increased risk for conversion to arthroplasty. There was no difference between failure/revision rates among the restorative approaches, yet cell-based approaches yielded a significantly increased risk for a return to the OR.

ABSTRACT

The purpose of this study is to describe the rate of return to the operating room (OR) following microfracture (MFX), autologous chondrocyte implantation (ACI), osteochondral autograft transplantation (OATS), and osteochondral allograft (OCA) procedures at 90 days, 1 year, and 2 years. Current Procedural Terminology codes for all patients undergoing MFX, ACI, OATS, and OCA were used to search a prospectively collected, commercially available private payer insurance company database from 2007 to 2011. Within 90 days, 1 year, and 2 years after surgery, the database was searched for the occurrence of these same patients undergoing knee diagnostic arthroscopy with biopsy, lysis of adhesions, synovectomy, arthroscopy for infection or lavage, arthroscopy for removal of loose bodies, chondroplasty, MFX, ACI, OATS, OCA, and/or knee arthroplasty. Descriptive statistical analysis and contingency table analysis were performed. A total of 47,207 cartilage procedures were performed from 2007 to 2011, including 43,576 MFX, 640 ACI, 386 open OATS, 997 arthroscopic OATS, 714 open OCA, and 894 arthroscopic OCA procedures. The weighted average reoperation rates for all procedures were 5.87% at 90 days, 11.94% at 1 year, and 14.90% at 2 years following the index cartilage surgery. At 2 years, patients who underwent MFX, ACI, OATS, OCA had reoperation rates of 14.65%, 29.69%, 8.82%, and 12.22%, respectively. There was a statistically significantly increased risk for ACI return to OR within all intervals (P < .0001); however, MFX had a greater risk factor (P < .0001) for conversion to arthroplasty. There was no difference in failure/revision rates between the restorative treatment options. With a large US commercial insurance database from 2007 to 2011, reparative procedures were favored for chondral injuries, but yielded an increased risk for conversion to arthroplasty. There was no difference in failure/revision rates between the restorative approaches, yet cell-based approaches yielded a significantly increased risk for a return to the OR.

Continue to: Symptomatic, full-thickness articular cartilage

 

Symptomatic, full-thickness articular cartilage defects in the knee are difficult to manage, particularly in the young, athletic patient population. Fortunately, a variety of cartilage repair (direct repair of the cartilage or those procedures which attempt to generate fibrocartilage) and restoration (those aimed at restoring hyaline cartilage) procedures are available, with encouraging short- and long-term clinical outcomes. After failure of nonoperative management, several surgical options are available for treating symptomatic focal chondral defects, including microfracture (MFX), autologous chondrocyte implantation (ACI), osteochondral autograft transplantation (OATS), and open and arthroscopic osteochondral allograft (OCA) transplantation procedures.^1,2 When appropriately indicated, each of these techniques has demonstrated good to excellent clinical outcomes with respect to reducing pain and improving function.^3-5

While major complications following cartilage surgery are uncommon, the need for reoperation following an index articular cartilage operation is poorly understood. Recently, McCormick and colleagues6 found that reoperation within the first 2 years following meniscus allograft transplantation (MAT) is associated with an increased likelihood of revision MAT or future arthroplasty. Given the association between early reoperation following meniscus restoration surgery and subsequent failure, an improved understanding of the epidemiology and implications of reoperations following cartilage restoration surgery is warranted. Further, in deciding which treatment option is best suited to a particular patient, the rate of return to the operating room (OR) should be taken into consideration, as this could potentially influence surgical decision-making as to which procedure to perform, especially in value-based care decision-making environments.

The purpose of this study is to describe the rate of return to the OR for knee procedures following cartilage restoration at intervals of 90 days, 1 year, and 2 years across a large-scale US patient database. The authors hypothesize that the rate of return to the OR following knee cartilage repair or restoration procedures will be under 20% during the first post-operative year, with increasing reoperation rates over time. A secondary hypothesis is that there will be no difference in reoperation rates according to sex, but that younger patients (those younger than 40 years) will have higher reoperation rates than older patients.

METHODS

We performed a retrospective analysis of a prospectively collected, large-scale, and commercially available private payer insurance company database (PearlDiver) from 2007 to 2011. The PearlDiver database is a Health Insurance Portability and Accountability Act (HIPAA) compliant, publicly available national database consisting of a collection of private payer records, with United Health Group representing the contributing health plan. The database has more than 30 million patient records and contains Current Procedural Terminology (CPT) and International Classification of Diseases, Ninth Revision (ICD-9) codes related to orthopedic procedures. From 2007 to 2011, the private payer database captured between 5.9 million and 6.2 million patients per year.

Our search was based on the CPT codes for MFX (29879), ACI (27412), OATS (29866, 29867), and OCA (27415, 27416). Return to the OR for revision surgery for the above-mentioned procedures was classified as patients with a diagnosis of diagnostic arthroscopy with biopsy (CPT 29870), lysis of adhesions (CPT 29884), synovectomy (29875, 29876), arthroscopy for infection or lavage (CPT 29871), arthroscopy for removal of loose bodies (29874), chondroplasty (29877), unicompartmental knee arthroplasty (27446), total knee arthroplasty (27447), and/or patellar arthroplasty (27438). Patient records were followed for reoperations occurring within 90 days, 1 year, and 2 years after the index cartilage procedure. All data were compared based on patient age and sex.

Table 1. Breakdown of MFX, ACI, OATS, and OCA Procedures by Sex
	MFX	ACI	Open OATS	Arthroscopic OATS	Open OCA	Arthroscopic OCA
Females	20,589	276	167	401	275	350
Males	22,987	364	219	596	439	544
Total	43,576	640	386	997	714	894

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation.

Continue to: Statistical analysis...

 

STATISTICAL ANALYSIS

Statistical analysis of this study was primarily descriptive to demonstrate the incidence for each code at each time interval. One-way analysis of variance, Chi-square analysis, and contingency tables were used to compare the incidence of each type of procedure throughout the various time intervals. A P-value of < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS v.20 (International Business Machines).

RESULTS

A total of 47,207 cartilage procedures were performed from 2007 to 2011, including 43,576 MFX (92.3%) 640 ACI (1.4%), 386 open OATS (0.82%), 997 arthroscopic OATS (2.11%), 714 open OCA (1.51%), and 894 arthroscopic OCA (1.89%) procedures. A summary of the procedures performed, broken down by age and sex, is provided in Tables 1 and 2. A total of 25,149 male patients (53.3%) underwent surgical procedures compared to 22,058 female patients (46.7%). For each category of procedure (MFX, ACI, OATS, OCA), there was a significantly higher proportion of males than females undergoing surgery (P < .0001 for all). Surgical treatment with MFX was consistently the most frequently performed surgery across all age groups (92.31%), while cell-based therapy with ACI was the least frequently performed procedure across all age ranges (1.36%). Restorative OATS and OCA techniques were performed with the greatest frequency in the 15-year-old to 19-year-old age group, but were not utilized in patients over 64 years of age (Table 2).

Table 2. Breakdown of MFX, ACI, OATS, and OCA Procedures by Age
Age (y)	MFX	ACI	OATS	OCA
10 to 14	572	22	74	47
15 to 19	1984	83	254	235
20 to 24	1468	54	140	144
25 to 29	1787	74	152	176
30 to 34	2824	114	152	204
35 to 39	4237	96	153	210
40 to 44	5441	103	166	217
45 to 49	7126	57	149	180
50 to 54	7004	25	83	140
55 to 59	6410	12	40	40
60 to 64	4409	0	20	15
65 to 69	269	0	0	0
70 to 74	45	0	0	0
Total	43,576	640	1383	1608

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation.

A summary of all reoperation data is provided in Tables 3 to 7 and Figures 1 and 2. The weighted average reoperation rates for all procedures were 5.87% at 90 days, 11.94% at 1 year, and 14.90% at 2 years following the index cartilage surgery. Patients who underwent MFX had reoperation rates of 6.05% at 90 days, 11.80% at 1 year, and 14.65% at 2 years. Patients who underwent ACI had reoperation rates of 4.53% at 90 days, 23.28% at 1 year, and 29.69% at 2 years. Patients who had open and arthroscopic OATS had reoperation rates of 3.122% and 5.12% at 90 days, 6.74% and 8.53% at 1 year, and 7.51% and 10.13% at 2 years, respectively. Patients who underwent open and arthroscopic OCA had reoperation rates of 2.52% and 3.91% at 90 days, 7.14% and 6.60% at 1 year, and 13.59% and 10.85% at 2 years (Table 3). There was a statistically significantly increased risk for reoperation following ACI within all intervals compared to all other surgical techniques (P < .0001); however, MFX had a greater risk factor (P < .0001) for conversion to arthroplasty at 6.70%. There was no significant difference between failure rates (revision OATS/OCA or conversion to arthroplasty) between the restorative treatment options, with 14 failures for OATS (9.52% of reoperations at 2 years) compared to 22 failures for OCA (12.7% of reoperations at 2 years, P = .358). Among the entire cohort of cartilage surgery patients, arthroscopic chondroplasty was the most frequent procedure performed at the time of reoperation at all time points assessed, notably accounting for 33.08% of reoperations 2 years following microfracture, 51.58% of reoperations at 2 years following ACI, 53.06% of reoperations at 2 years following OATS, and 54.07% of reoperations at 2 years following OCA (Figure 3, Tables 4–7).

Table 3. Comparison of Return to OR Following MFX, ACI, OCA, and OATS
Procedure	Total No. of Cases in Study Period	No. of Reoperations at 90 Days	Return to OR Rate at 90 Days	No. of Reoperations at 1 Year	Return to OR Rate at 1 Year	No. of Reoperations at 2 Years	Return to OR Rate at 2 Years
MFX	43,576	2636	6.05%	5142	11.80%	6385	14.65%
ACI	640	29	4.53%	149	23.28%	190	29.69%
Open OATS	386	12	3.12%	26	6.74%	29	7.51%
Arthroscopic OATS	997	51	5.12%	85	8.53%	101	10.13%
Open OCA	714	18	2.52%	51	7.14%	97	13.59%
Arthroscopic OCA	894	161	3.91%	59	6.60%	97	10.85%
Weighted average for all procedures			5.87%		11.94%		14.90%

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation; OR, operating room.

Table 4. Rate of Return to OR Following MFX (n = 43,574)
Procedure	CPT Code	90 Days	1 Year	2 Years
Knee arthroscopy	29870	54	122	162
Knee arthroscopic drainage and lavage	29871	84	102	104
Arthroscopic adhesions débridement	29874	300	468	549
Arthroscopic synovectomy	29875	324	528	611
Major arthroscopic synovectomy	29876	557	926	1087
Knee arthroscopic chondroplasty	29877	1063	1722	2112
Arthroscopic lysis of adhesions	29884	61	129	171
Patellar arthroplasty	27438	0	38	49
Medial or lateral knee arthroplasty	27446	51	242	328
Medial and lateral knee arthroplasty	27447	142	865	1212
Total		2636	5142	6385
Return to OR		6.05%	11.80%	14.65%

Abbreviations: CPT, Current Procedural Terminology; MFX, microfracture; OR, operating room.

Table 5. Rate of Return to OR Following ACI (n = 640)
Procedure	CPT Code	90 Daysa	1 Yeara	2 Yearsa
Revision ACI	27412	29	33	35
Knee arthroscopy	29870	-1	-1	-1
Knee arthroscopic drainage and lavage	29871	-1	-1	-1
Arthroscopic adhesions débridement	29874	0	-1	-1
Arthroscopic synovectomy	29875	-1	-1	-1
Major arthroscopic synovectomy	29876	-1	12	20
Knee arthroscopic chondroplasty	29877	-1	71	98
Arthroscopic lysis of adhesions	29884	-1	33	37
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	-1	-1
Medial and lateral knee arthroplasty	27447	0	-1	-1
Total		29	149	190
Return to OR		4.53%	23.28%	29.69%

^aA -1 denotes No. <11 within the PearlDiver database, and exact numbers are not reported due to patient privacy considerations.

Abbreviations: ACI, autologous chondrocyte implantation; CPT, Current Procedural Terminology; OR, operating room.

Table 6. Rate of Return to OR Following OATS (n = 1320)
Procedure	CPT Code	90 Days	1 Year	2 Years
Knee arthroscopy	29870	0	0	0
Knee arthroscopic drainage and lavage	29871	0	0	0
Arthroscopic adhesions débridement	29874	0	12	13
Arthroscopic synovectomy	29875	0	0	14
Major arthroscopic synovectomy	29876	16	25	28
Knee arthroscopic chondroplasty	29877	17	58	78
Arthroscopic lysis of adhesions	29884	0	0	0
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	0	0
Medial and lateral knee arthroplasty	27447	0	0	14
Total		33	95	147
Return to OR		2.50%	7.20%	11.14%

Abbreviations: CPT, Current Procedural Terminology; OATS, osteochondral autograft transplantation; OR, operating room.

Table 7. Rate of Return to OR Following OCA Transplantation (n = 1531)
Procedure	CPT Code	90 Days	1 Year	2 Year
Knee arthroscopy	29870	0	0	0
Knee arthroscopic drainage and lavage	29871	0	0	0
Arthroscopic adhesions débridement	29874	0	15	19
Arthroscopic synovectomy	29875	0	0	0
Major arthroscopic synovectomy	29876	0	20	38
Knee arthroscopic chondroplasty	29877	22	59	93
Arthroscopic lysis of adhesions	29884	0	0	0
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	0	0
Medial and lateral knee arthroplasty	27447	0	0	22
Total		22	94	172
Return to OR		1.44%	6.14%	11.23%

Abbreviations: CPT, Current Procedural Terminology; OCA, osteochondral allograft; OR, operating room.

Continue to: Discussion...

 

 

DISCUSSION

The principle findings of this study demonstrate that there is an overall reoperation rate of 14.90% at 2 years following cartilage repair/restoration surgery, with the highest reoperation rates following MFX at 90 days, and ACI at both 1 year and 2 years following the index procedure. Also, patients undergoing index MFX as the index procedure have the highest risk for conversion to arthroplasty, reoperation rates for all cartilage surgeries increase over time, and arthroscopic chondroplasty is the most frequent procedure performed at the time of reoperation.

The management of symptomatic articular cartilage knee pathology is extremely challenging. With improvements in surgical technique, instrumentation, and clinical decision-making, indications are constantly evolving. Techniques that may work for “small” defects, though there is some debate as to what constitutes a “small” defect, are not necessarily going to be successful for larger defects, and this certainly varies depending on where the defect is located within the knee joint (distal femur vs patella vs trochlea, etc.). Recently, in a 2015 analysis of 3 level I or II studies, Miller and colleagues⁷ demonstrated both MFX and OATS to be viable, cost-effective, first-line treatment options for articular cartilage injuries, with similar clinical outcomes at 8.7 years. The authors noted cumulative reoperation rates of 29% among patients undergoing MFX compared to 13% among patients undergoing OATS. While ACI and OCA procedures were not included in their study, the reported reoperation rates of 29% following MFX and 13% following OATS at nearly 10 years suggest a possible increased need for reoperation following MFX over time (approximately 15% at 2 years in our study) and a stable rate of reoperation following OATS (approximately 11% at 2 years in our study). This finding is significant, as one of the goals with these procedures is to deliver effective, long-lasting pain relief and restoration of function. Interestingly, in this study, restorative OATS and OCA techniques were performed with the greatest frequency in the 15-year-old to 19-year-old age group, but were not performed in patients older than 64 years. This may be explained by the higher prevalence of acute traumatic injuries and osteochondritis dissecans diagnoses in younger patients compared with older patients, as these diagnoses are more often indicated to undergo restorative procedures as opposed to marrow stimulation.

In a 2016 systematic review of 20 studies incorporating 1117 patients, Campbell and colleagues⁸ assessed return-to-play rates following MFX, ACI, OATS, and OCA. The authors noted that return to sport (RTS) rates were greatest following OATS (89%), followed by OCA (88%), ACI (84%), and MFX (75%). Positive prognostic factors for RTS included younger age, shorter duration of preoperative symptoms, no history of prior ipsilateral knee surgery, and smaller chondral defects. Reoperation rates between the 4 techniques were not statistically compared in their study. Interestingly, in 2013, Chalmers and colleagues⁹ conducted a separate systematic review of 20 studies comprising 1375 patients undergoing MFX, ACI, or OATS. In their study, the authors found significant advantages following ACI and OATS compared to MFX with respect to patient-reported outcome scores but noted significantly faster RTS rates with MFX. Reoperation rates were noted to be similar between the 3 procedures (25% for ACI, 21% for MFX, and 28% for OATS) at an average 3.7 years following the index procedure. When considering these 2 systematic reviews together, despite a faster RTS rate following MFX, a greater proportion of patients seem to be able to RTS over time following other procedures such as OATS, OCA, and ACI. Unfortunately, these reviews do not provide insight as to the role, if any, of reoperation on return to play rates nor on overall clinical outcome scores on patients undergoing articular cartilage surgery. However, this information is valuable when counseling athletes who are in season and would like to RTS as soon as possible as opposed to those who do not have tight time constraints for when they need to RTS.

Regardless of the cartilage technique chosen, the goals of surgery remain similar—to reduce pain and improve function. For athletes, the ultimate goal is to return to the same level of play that the athlete was able to achieve prior to injury. Certainly, the need for reoperation following a cartilage surgery has implications on pain, function, and ability to RTS. Our review of nearly 50,000 cartilage surgeries demonstrates that reoperations following cartilage repair surgery are not uncommon, with a rate of 14.90% at 2 years, and that while reoperation rates are the highest following ACI, the rate of conversion to knee arthroplasty is highest following MFX. Due to the limitations of the PearlDiver database, it is not possible to determine the clinical outcomes of patients undergoing reoperation following cartilage surgery, but certainly, given these data, reoperation is clearly not necessarily indicative of clinical failure. This is highlighted by the fact that the most common procedure performed at the time of reoperation is arthroscopic chondroplasty, which, despite being an additional surgical procedure, may be acceptable for patients who wish to RTS, particularly in the setting of an index ACI in which there may be graft hypertrophy. Ideally, additional studies incorporating a cost-effectiveness analysis of each of the procedures, incorporating reoperation rates as well as patient-reported clinical outcomes, would be helpful to truly determine the patient and societal implications of reoperation following cartilage repair/restoration.

Many of the advantages and disadvantages of the described cartilage repair/restoration procedures have been well described.^10-17 Microfracture is the most commonly utilized first-line repair/restoration option for small articular cartilage lesions, mainly due to its low cost, low morbidity, and relatively low level of difficulty.18 Despite these advantages, MFX is not without limitations, and the need for revision cartilage restoration and/or conversion to arthroplasty is concerning. In 2013, Salzmann and colleagues19 evaluated a cohort of 454 patients undergoing MFX for a symptomatic knee defect and noted a reoperation rate of 26.9% (n = 123) within 2 years of the index surgery, with risk factors for reoperation noted to include an increased number of pre-MFX ipsilateral knee surgeries, patellofemoral lesions, smoking, and lower preoperative numeric analog scale scores. The definition of reoperation in their study is unfortunately not described, and thus the extent of reoperation (arthroscopy to arthroplasty) is unclear. In a 2009 systematic review of 3122 patients (28 studies) undergoing MFX conducted by Mithoefer and colleagues,²⁰ revision rates were noted to range from 2% to 31% depending on the study analyzed, with increasing revision rates after 2 years. Unfortunately, the heterogeneity of the included studies makes it difficult to determine which patients tend to fail over time.

Continue to: OATS...

 

OATS is a promising cartilage restoration technique indicated for treatment of patients with large, uncontained chondral lesions, and/or lesions with both bone and cartilage loss.¹ OCA is similar to OATS but uses allograft tissue instead of autograft tissue and is typically considered a viable treatment option in larger lesions (>2 cm²).²¹ Cell-based ACI therapy has evolved substantially over the past decade and is now available as a third-generation model utilizing biodegradable 3-dimensional scaffolds seeded with chondrocytes. Reoperation rates following ACI can often be higher than those following other cartilage treatments, particularly given the known complication of graft hypertrophy and/or delamination. Harris and colleagues²² conducted a systematic review of 5276 subjects undergoing ACI (all generations), noting an overall reoperation rate of 33%, but a failure rate of 5.8% at an average of 22 months following ACI. Risk factors for reoperation included periosteal-based ACI as well as open (vs arthroscopic) ACI. In this study, we found a modestly lower return to OR rate of 29.69% at 2 years.

When the outcomes of patients undergoing OATS or OCA are compared to those of patients undergoing MFX or ACI, it can be difficult to interpret the results, as the indications for performing these procedures tend to be very different. Further, the reasons for reoperation, as well as the procedures performed at the time of reoperation, are often poorly described, making it difficult to truly quantify the risk of reoperation and the implications of reoperation for patients undergoing any of these index cartilage procedures.

Overall, in this database, the return to the OR rate approaches 15% at 2 years following cartilage surgery, with cell-based therapy demonstrating higher reoperation rates at 2 years, without the risk of conversion to arthroplasty. Reoperation rates appear to stabilize at 1 year following surgery and consist mostly of minor arthroscopic procedures. These findings can help surgeons counsel patients as to the rate and type of reoperations that can be expected following cartilage surgery. Additional research incorporating patient-reported outcomes and patient-specific risk factors are needed to complement these data as to the impact of reoperations on overall clinical outcomes. Further, studies incorporating 90-day, 1-year, and 2-year costs associated with cartilage surgery will help to determine which index procedure is the most cost effective over the short- and long-term.

LIMITATIONS

This study is not without limitations. The PearlDiver database is reliant upon accurate CPT and ICD-9 coding, which creates a potential for a reporting bias. The overall reliability of the analyses is dependent on the quality of the available data, which, as noted in previous PearlDiver studies,^18,23-28 may include inaccurate billing codes, miscoding, and/or non-coding by physicians as potential sources of error. At the time of this study, the PearlDiver database did not provide consistent data points on laterality, and thus it is possible that the reported rates of reoperation overestimate the true reoperation rate following a given procedure. Fortunately, the reoperation rates for each procedure analyzed in this database study are consistent with those previously presented in the literature. In addition, it is not uncommon for patients receiving one of these procedures to have previously been treated with one of the others. Due to the inherent limitations of the PearlDiver database, this study did not investigate concomitant procedures performed along with the index procedure, nor did it investigate confounding factors such as comorbidities. The PearlDiver database does not provide data on defect size, location within the knee, concomitant pathologies (eg, meniscus tear), prior surgeries, or patient comorbidities, and while important, these factors cannot be accounted for in our analysis. The inability to account for these important factors, particularly concomitant diagnoses, procedures, and lesion size/location, represents an important limitation of this study, as this is a source of selection bias and may influence the need for reoperation in a given patient. Despite these limitations, the results of this study are supported by previous and current literature. In addition, the PearlDiver database, as a HIPAA-compliant database, does not report exact numbers when the value of the outcome of interest is between 0 and 10, which prohibits analysis of any cartilage procedure performed in a cohort of patients greater than 1 and less than 11. Finally, while not necessarily a limitation, it should be noted that CPT 29879 is not specific for microfracture, as the code also includes abrasion arthroplasty and drilling. Due to the limitations of the methodology of searching the database for this code, it is unclear as to how many patients underwent actual microfracture vs abrasion arthroplasty.

CONCLUSION

Within a large US commercial insurance database from 2007 to 2011, reparative procedures were favored for chondral injuries, but yielded an increased risk for conversion to arthroplasty. There was no difference between failure/revision rates among the restorative approaches, yet cell-based approaches yielded a significantly increased risk for a return to the OR.

ABSTRACT

The purpose of this study is to describe the rate of return to the operating room (OR) following microfracture (MFX), autologous chondrocyte implantation (ACI), osteochondral autograft transplantation (OATS), and osteochondral allograft (OCA) procedures at 90 days, 1 year, and 2 years. Current Procedural Terminology codes for all patients undergoing MFX, ACI, OATS, and OCA were used to search a prospectively collected, commercially available private payer insurance company database from 2007 to 2011. Within 90 days, 1 year, and 2 years after surgery, the database was searched for the occurrence of these same patients undergoing knee diagnostic arthroscopy with biopsy, lysis of adhesions, synovectomy, arthroscopy for infection or lavage, arthroscopy for removal of loose bodies, chondroplasty, MFX, ACI, OATS, OCA, and/or knee arthroplasty. Descriptive statistical analysis and contingency table analysis were performed. A total of 47,207 cartilage procedures were performed from 2007 to 2011, including 43,576 MFX, 640 ACI, 386 open OATS, 997 arthroscopic OATS, 714 open OCA, and 894 arthroscopic OCA procedures. The weighted average reoperation rates for all procedures were 5.87% at 90 days, 11.94% at 1 year, and 14.90% at 2 years following the index cartilage surgery. At 2 years, patients who underwent MFX, ACI, OATS, OCA had reoperation rates of 14.65%, 29.69%, 8.82%, and 12.22%, respectively. There was a statistically significantly increased risk for ACI return to OR within all intervals (P < .0001); however, MFX had a greater risk factor (P < .0001) for conversion to arthroplasty. There was no difference in failure/revision rates between the restorative treatment options. With a large US commercial insurance database from 2007 to 2011, reparative procedures were favored for chondral injuries, but yielded an increased risk for conversion to arthroplasty. There was no difference in failure/revision rates between the restorative approaches, yet cell-based approaches yielded a significantly increased risk for a return to the OR.

Continue to: Symptomatic, full-thickness articular cartilage

 

Symptomatic, full-thickness articular cartilage defects in the knee are difficult to manage, particularly in the young, athletic patient population. Fortunately, a variety of cartilage repair (direct repair of the cartilage or those procedures which attempt to generate fibrocartilage) and restoration (those aimed at restoring hyaline cartilage) procedures are available, with encouraging short- and long-term clinical outcomes. After failure of nonoperative management, several surgical options are available for treating symptomatic focal chondral defects, including microfracture (MFX), autologous chondrocyte implantation (ACI), osteochondral autograft transplantation (OATS), and open and arthroscopic osteochondral allograft (OCA) transplantation procedures.^1,2 When appropriately indicated, each of these techniques has demonstrated good to excellent clinical outcomes with respect to reducing pain and improving function.^3-5

While major complications following cartilage surgery are uncommon, the need for reoperation following an index articular cartilage operation is poorly understood. Recently, McCormick and colleagues6 found that reoperation within the first 2 years following meniscus allograft transplantation (MAT) is associated with an increased likelihood of revision MAT or future arthroplasty. Given the association between early reoperation following meniscus restoration surgery and subsequent failure, an improved understanding of the epidemiology and implications of reoperations following cartilage restoration surgery is warranted. Further, in deciding which treatment option is best suited to a particular patient, the rate of return to the operating room (OR) should be taken into consideration, as this could potentially influence surgical decision-making as to which procedure to perform, especially in value-based care decision-making environments.

The purpose of this study is to describe the rate of return to the OR for knee procedures following cartilage restoration at intervals of 90 days, 1 year, and 2 years across a large-scale US patient database. The authors hypothesize that the rate of return to the OR following knee cartilage repair or restoration procedures will be under 20% during the first post-operative year, with increasing reoperation rates over time. A secondary hypothesis is that there will be no difference in reoperation rates according to sex, but that younger patients (those younger than 40 years) will have higher reoperation rates than older patients.

METHODS

We performed a retrospective analysis of a prospectively collected, large-scale, and commercially available private payer insurance company database (PearlDiver) from 2007 to 2011. The PearlDiver database is a Health Insurance Portability and Accountability Act (HIPAA) compliant, publicly available national database consisting of a collection of private payer records, with United Health Group representing the contributing health plan. The database has more than 30 million patient records and contains Current Procedural Terminology (CPT) and International Classification of Diseases, Ninth Revision (ICD-9) codes related to orthopedic procedures. From 2007 to 2011, the private payer database captured between 5.9 million and 6.2 million patients per year.

Our search was based on the CPT codes for MFX (29879), ACI (27412), OATS (29866, 29867), and OCA (27415, 27416). Return to the OR for revision surgery for the above-mentioned procedures was classified as patients with a diagnosis of diagnostic arthroscopy with biopsy (CPT 29870), lysis of adhesions (CPT 29884), synovectomy (29875, 29876), arthroscopy for infection or lavage (CPT 29871), arthroscopy for removal of loose bodies (29874), chondroplasty (29877), unicompartmental knee arthroplasty (27446), total knee arthroplasty (27447), and/or patellar arthroplasty (27438). Patient records were followed for reoperations occurring within 90 days, 1 year, and 2 years after the index cartilage procedure. All data were compared based on patient age and sex.

Table 1. Breakdown of MFX, ACI, OATS, and OCA Procedures by Sex
	MFX	ACI	Open OATS	Arthroscopic OATS	Open OCA	Arthroscopic OCA
Females	20,589	276	167	401	275	350
Males	22,987	364	219	596	439	544
Total	43,576	640	386	997	714	894

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation.

Continue to: Statistical analysis...

 

STATISTICAL ANALYSIS

Statistical analysis of this study was primarily descriptive to demonstrate the incidence for each code at each time interval. One-way analysis of variance, Chi-square analysis, and contingency tables were used to compare the incidence of each type of procedure throughout the various time intervals. A P-value of < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS v.20 (International Business Machines).

RESULTS

A total of 47,207 cartilage procedures were performed from 2007 to 2011, including 43,576 MFX (92.3%) 640 ACI (1.4%), 386 open OATS (0.82%), 997 arthroscopic OATS (2.11%), 714 open OCA (1.51%), and 894 arthroscopic OCA (1.89%) procedures. A summary of the procedures performed, broken down by age and sex, is provided in Tables 1 and 2. A total of 25,149 male patients (53.3%) underwent surgical procedures compared to 22,058 female patients (46.7%). For each category of procedure (MFX, ACI, OATS, OCA), there was a significantly higher proportion of males than females undergoing surgery (P < .0001 for all). Surgical treatment with MFX was consistently the most frequently performed surgery across all age groups (92.31%), while cell-based therapy with ACI was the least frequently performed procedure across all age ranges (1.36%). Restorative OATS and OCA techniques were performed with the greatest frequency in the 15-year-old to 19-year-old age group, but were not utilized in patients over 64 years of age (Table 2).

Table 2. Breakdown of MFX, ACI, OATS, and OCA Procedures by Age
Age (y)	MFX	ACI	OATS	OCA
10 to 14	572	22	74	47
15 to 19	1984	83	254	235
20 to 24	1468	54	140	144
25 to 29	1787	74	152	176
30 to 34	2824	114	152	204
35 to 39	4237	96	153	210
40 to 44	5441	103	166	217
45 to 49	7126	57	149	180
50 to 54	7004	25	83	140
55 to 59	6410	12	40	40
60 to 64	4409	0	20	15
65 to 69	269	0	0	0
70 to 74	45	0	0	0
Total	43,576	640	1383	1608

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation.

A summary of all reoperation data is provided in Tables 3 to 7 and Figures 1 and 2. The weighted average reoperation rates for all procedures were 5.87% at 90 days, 11.94% at 1 year, and 14.90% at 2 years following the index cartilage surgery. Patients who underwent MFX had reoperation rates of 6.05% at 90 days, 11.80% at 1 year, and 14.65% at 2 years. Patients who underwent ACI had reoperation rates of 4.53% at 90 days, 23.28% at 1 year, and 29.69% at 2 years. Patients who had open and arthroscopic OATS had reoperation rates of 3.122% and 5.12% at 90 days, 6.74% and 8.53% at 1 year, and 7.51% and 10.13% at 2 years, respectively. Patients who underwent open and arthroscopic OCA had reoperation rates of 2.52% and 3.91% at 90 days, 7.14% and 6.60% at 1 year, and 13.59% and 10.85% at 2 years (Table 3). There was a statistically significantly increased risk for reoperation following ACI within all intervals compared to all other surgical techniques (P < .0001); however, MFX had a greater risk factor (P < .0001) for conversion to arthroplasty at 6.70%. There was no significant difference between failure rates (revision OATS/OCA or conversion to arthroplasty) between the restorative treatment options, with 14 failures for OATS (9.52% of reoperations at 2 years) compared to 22 failures for OCA (12.7% of reoperations at 2 years, P = .358). Among the entire cohort of cartilage surgery patients, arthroscopic chondroplasty was the most frequent procedure performed at the time of reoperation at all time points assessed, notably accounting for 33.08% of reoperations 2 years following microfracture, 51.58% of reoperations at 2 years following ACI, 53.06% of reoperations at 2 years following OATS, and 54.07% of reoperations at 2 years following OCA (Figure 3, Tables 4–7).

Table 3. Comparison of Return to OR Following MFX, ACI, OCA, and OATS
Procedure	Total No. of Cases in Study Period	No. of Reoperations at 90 Days	Return to OR Rate at 90 Days	No. of Reoperations at 1 Year	Return to OR Rate at 1 Year	No. of Reoperations at 2 Years	Return to OR Rate at 2 Years
MFX	43,576	2636	6.05%	5142	11.80%	6385	14.65%
ACI	640	29	4.53%	149	23.28%	190	29.69%
Open OATS	386	12	3.12%	26	6.74%	29	7.51%
Arthroscopic OATS	997	51	5.12%	85	8.53%	101	10.13%
Open OCA	714	18	2.52%	51	7.14%	97	13.59%
Arthroscopic OCA	894	161	3.91%	59	6.60%	97	10.85%
Weighted average for all procedures			5.87%		11.94%		14.90%

Abbreviations: ACI, autologous chondrocyte implantation; MFX, microfracture; OCA, osteochondral allograft; OATS, osteochondral autograft transplantation; OR, operating room.

Table 4. Rate of Return to OR Following MFX (n = 43,574)
Procedure	CPT Code	90 Days	1 Year	2 Years
Knee arthroscopy	29870	54	122	162
Knee arthroscopic drainage and lavage	29871	84	102	104
Arthroscopic adhesions débridement	29874	300	468	549
Arthroscopic synovectomy	29875	324	528	611
Major arthroscopic synovectomy	29876	557	926	1087
Knee arthroscopic chondroplasty	29877	1063	1722	2112
Arthroscopic lysis of adhesions	29884	61	129	171
Patellar arthroplasty	27438	0	38	49
Medial or lateral knee arthroplasty	27446	51	242	328
Medial and lateral knee arthroplasty	27447	142	865	1212
Total		2636	5142	6385
Return to OR		6.05%	11.80%	14.65%

Abbreviations: CPT, Current Procedural Terminology; MFX, microfracture; OR, operating room.

Table 5. Rate of Return to OR Following ACI (n = 640)
Procedure	CPT Code	90 Daysa	1 Yeara	2 Yearsa
Revision ACI	27412	29	33	35
Knee arthroscopy	29870	-1	-1	-1
Knee arthroscopic drainage and lavage	29871	-1	-1	-1
Arthroscopic adhesions débridement	29874	0	-1	-1
Arthroscopic synovectomy	29875	-1	-1	-1
Major arthroscopic synovectomy	29876	-1	12	20
Knee arthroscopic chondroplasty	29877	-1	71	98
Arthroscopic lysis of adhesions	29884	-1	33	37
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	-1	-1
Medial and lateral knee arthroplasty	27447	0	-1	-1
Total		29	149	190
Return to OR		4.53%	23.28%	29.69%

^aA -1 denotes No. <11 within the PearlDiver database, and exact numbers are not reported due to patient privacy considerations.

Abbreviations: ACI, autologous chondrocyte implantation; CPT, Current Procedural Terminology; OR, operating room.

Table 6. Rate of Return to OR Following OATS (n = 1320)
Procedure	CPT Code	90 Days	1 Year	2 Years
Knee arthroscopy	29870	0	0	0
Knee arthroscopic drainage and lavage	29871	0	0	0
Arthroscopic adhesions débridement	29874	0	12	13
Arthroscopic synovectomy	29875	0	0	14
Major arthroscopic synovectomy	29876	16	25	28
Knee arthroscopic chondroplasty	29877	17	58	78
Arthroscopic lysis of adhesions	29884	0	0	0
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	0	0
Medial and lateral knee arthroplasty	27447	0	0	14
Total		33	95	147
Return to OR		2.50%	7.20%	11.14%

Abbreviations: CPT, Current Procedural Terminology; OATS, osteochondral autograft transplantation; OR, operating room.

Table 7. Rate of Return to OR Following OCA Transplantation (n = 1531)
Procedure	CPT Code	90 Days	1 Year	2 Year
Knee arthroscopy	29870	0	0	0
Knee arthroscopic drainage and lavage	29871	0	0	0
Arthroscopic adhesions débridement	29874	0	15	19
Arthroscopic synovectomy	29875	0	0	0
Major arthroscopic synovectomy	29876	0	20	38
Knee arthroscopic chondroplasty	29877	22	59	93
Arthroscopic lysis of adhesions	29884	0	0	0
Patellar arthroplasty	27438	0	0	0
Medial or lateral knee arthroplasty	27446	0	0	0
Medial and lateral knee arthroplasty	27447	0	0	22
Total		22	94	172
Return to OR		1.44%	6.14%	11.23%

Abbreviations: CPT, Current Procedural Terminology; OCA, osteochondral allograft; OR, operating room.

Continue to: Discussion...

 

 

DISCUSSION

The principle findings of this study demonstrate that there is an overall reoperation rate of 14.90% at 2 years following cartilage repair/restoration surgery, with the highest reoperation rates following MFX at 90 days, and ACI at both 1 year and 2 years following the index procedure. Also, patients undergoing index MFX as the index procedure have the highest risk for conversion to arthroplasty, reoperation rates for all cartilage surgeries increase over time, and arthroscopic chondroplasty is the most frequent procedure performed at the time of reoperation.

The management of symptomatic articular cartilage knee pathology is extremely challenging. With improvements in surgical technique, instrumentation, and clinical decision-making, indications are constantly evolving. Techniques that may work for “small” defects, though there is some debate as to what constitutes a “small” defect, are not necessarily going to be successful for larger defects, and this certainly varies depending on where the defect is located within the knee joint (distal femur vs patella vs trochlea, etc.). Recently, in a 2015 analysis of 3 level I or II studies, Miller and colleagues⁷ demonstrated both MFX and OATS to be viable, cost-effective, first-line treatment options for articular cartilage injuries, with similar clinical outcomes at 8.7 years. The authors noted cumulative reoperation rates of 29% among patients undergoing MFX compared to 13% among patients undergoing OATS. While ACI and OCA procedures were not included in their study, the reported reoperation rates of 29% following MFX and 13% following OATS at nearly 10 years suggest a possible increased need for reoperation following MFX over time (approximately 15% at 2 years in our study) and a stable rate of reoperation following OATS (approximately 11% at 2 years in our study). This finding is significant, as one of the goals with these procedures is to deliver effective, long-lasting pain relief and restoration of function. Interestingly, in this study, restorative OATS and OCA techniques were performed with the greatest frequency in the 15-year-old to 19-year-old age group, but were not performed in patients older than 64 years. This may be explained by the higher prevalence of acute traumatic injuries and osteochondritis dissecans diagnoses in younger patients compared with older patients, as these diagnoses are more often indicated to undergo restorative procedures as opposed to marrow stimulation.

In a 2016 systematic review of 20 studies incorporating 1117 patients, Campbell and colleagues⁸ assessed return-to-play rates following MFX, ACI, OATS, and OCA. The authors noted that return to sport (RTS) rates were greatest following OATS (89%), followed by OCA (88%), ACI (84%), and MFX (75%). Positive prognostic factors for RTS included younger age, shorter duration of preoperative symptoms, no history of prior ipsilateral knee surgery, and smaller chondral defects. Reoperation rates between the 4 techniques were not statistically compared in their study. Interestingly, in 2013, Chalmers and colleagues⁹ conducted a separate systematic review of 20 studies comprising 1375 patients undergoing MFX, ACI, or OATS. In their study, the authors found significant advantages following ACI and OATS compared to MFX with respect to patient-reported outcome scores but noted significantly faster RTS rates with MFX. Reoperation rates were noted to be similar between the 3 procedures (25% for ACI, 21% for MFX, and 28% for OATS) at an average 3.7 years following the index procedure. When considering these 2 systematic reviews together, despite a faster RTS rate following MFX, a greater proportion of patients seem to be able to RTS over time following other procedures such as OATS, OCA, and ACI. Unfortunately, these reviews do not provide insight as to the role, if any, of reoperation on return to play rates nor on overall clinical outcome scores on patients undergoing articular cartilage surgery. However, this information is valuable when counseling athletes who are in season and would like to RTS as soon as possible as opposed to those who do not have tight time constraints for when they need to RTS.

Regardless of the cartilage technique chosen, the goals of surgery remain similar—to reduce pain and improve function. For athletes, the ultimate goal is to return to the same level of play that the athlete was able to achieve prior to injury. Certainly, the need for reoperation following a cartilage surgery has implications on pain, function, and ability to RTS. Our review of nearly 50,000 cartilage surgeries demonstrates that reoperations following cartilage repair surgery are not uncommon, with a rate of 14.90% at 2 years, and that while reoperation rates are the highest following ACI, the rate of conversion to knee arthroplasty is highest following MFX. Due to the limitations of the PearlDiver database, it is not possible to determine the clinical outcomes of patients undergoing reoperation following cartilage surgery, but certainly, given these data, reoperation is clearly not necessarily indicative of clinical failure. This is highlighted by the fact that the most common procedure performed at the time of reoperation is arthroscopic chondroplasty, which, despite being an additional surgical procedure, may be acceptable for patients who wish to RTS, particularly in the setting of an index ACI in which there may be graft hypertrophy. Ideally, additional studies incorporating a cost-effectiveness analysis of each of the procedures, incorporating reoperation rates as well as patient-reported clinical outcomes, would be helpful to truly determine the patient and societal implications of reoperation following cartilage repair/restoration.

Many of the advantages and disadvantages of the described cartilage repair/restoration procedures have been well described.^10-17 Microfracture is the most commonly utilized first-line repair/restoration option for small articular cartilage lesions, mainly due to its low cost, low morbidity, and relatively low level of difficulty.18 Despite these advantages, MFX is not without limitations, and the need for revision cartilage restoration and/or conversion to arthroplasty is concerning. In 2013, Salzmann and colleagues19 evaluated a cohort of 454 patients undergoing MFX for a symptomatic knee defect and noted a reoperation rate of 26.9% (n = 123) within 2 years of the index surgery, with risk factors for reoperation noted to include an increased number of pre-MFX ipsilateral knee surgeries, patellofemoral lesions, smoking, and lower preoperative numeric analog scale scores. The definition of reoperation in their study is unfortunately not described, and thus the extent of reoperation (arthroscopy to arthroplasty) is unclear. In a 2009 systematic review of 3122 patients (28 studies) undergoing MFX conducted by Mithoefer and colleagues,²⁰ revision rates were noted to range from 2% to 31% depending on the study analyzed, with increasing revision rates after 2 years. Unfortunately, the heterogeneity of the included studies makes it difficult to determine which patients tend to fail over time.

Continue to: OATS...

 

OATS is a promising cartilage restoration technique indicated for treatment of patients with large, uncontained chondral lesions, and/or lesions with both bone and cartilage loss.¹ OCA is similar to OATS but uses allograft tissue instead of autograft tissue and is typically considered a viable treatment option in larger lesions (>2 cm²).²¹ Cell-based ACI therapy has evolved substantially over the past decade and is now available as a third-generation model utilizing biodegradable 3-dimensional scaffolds seeded with chondrocytes. Reoperation rates following ACI can often be higher than those following other cartilage treatments, particularly given the known complication of graft hypertrophy and/or delamination. Harris and colleagues²² conducted a systematic review of 5276 subjects undergoing ACI (all generations), noting an overall reoperation rate of 33%, but a failure rate of 5.8% at an average of 22 months following ACI. Risk factors for reoperation included periosteal-based ACI as well as open (vs arthroscopic) ACI. In this study, we found a modestly lower return to OR rate of 29.69% at 2 years.

When the outcomes of patients undergoing OATS or OCA are compared to those of patients undergoing MFX or ACI, it can be difficult to interpret the results, as the indications for performing these procedures tend to be very different. Further, the reasons for reoperation, as well as the procedures performed at the time of reoperation, are often poorly described, making it difficult to truly quantify the risk of reoperation and the implications of reoperation for patients undergoing any of these index cartilage procedures.

Overall, in this database, the return to the OR rate approaches 15% at 2 years following cartilage surgery, with cell-based therapy demonstrating higher reoperation rates at 2 years, without the risk of conversion to arthroplasty. Reoperation rates appear to stabilize at 1 year following surgery and consist mostly of minor arthroscopic procedures. These findings can help surgeons counsel patients as to the rate and type of reoperations that can be expected following cartilage surgery. Additional research incorporating patient-reported outcomes and patient-specific risk factors are needed to complement these data as to the impact of reoperations on overall clinical outcomes. Further, studies incorporating 90-day, 1-year, and 2-year costs associated with cartilage surgery will help to determine which index procedure is the most cost effective over the short- and long-term.

LIMITATIONS

This study is not without limitations. The PearlDiver database is reliant upon accurate CPT and ICD-9 coding, which creates a potential for a reporting bias. The overall reliability of the analyses is dependent on the quality of the available data, which, as noted in previous PearlDiver studies,^18,23-28 may include inaccurate billing codes, miscoding, and/or non-coding by physicians as potential sources of error. At the time of this study, the PearlDiver database did not provide consistent data points on laterality, and thus it is possible that the reported rates of reoperation overestimate the true reoperation rate following a given procedure. Fortunately, the reoperation rates for each procedure analyzed in this database study are consistent with those previously presented in the literature. In addition, it is not uncommon for patients receiving one of these procedures to have previously been treated with one of the others. Due to the inherent limitations of the PearlDiver database, this study did not investigate concomitant procedures performed along with the index procedure, nor did it investigate confounding factors such as comorbidities. The PearlDiver database does not provide data on defect size, location within the knee, concomitant pathologies (eg, meniscus tear), prior surgeries, or patient comorbidities, and while important, these factors cannot be accounted for in our analysis. The inability to account for these important factors, particularly concomitant diagnoses, procedures, and lesion size/location, represents an important limitation of this study, as this is a source of selection bias and may influence the need for reoperation in a given patient. Despite these limitations, the results of this study are supported by previous and current literature. In addition, the PearlDiver database, as a HIPAA-compliant database, does not report exact numbers when the value of the outcome of interest is between 0 and 10, which prohibits analysis of any cartilage procedure performed in a cohort of patients greater than 1 and less than 11. Finally, while not necessarily a limitation, it should be noted that CPT 29879 is not specific for microfracture, as the code also includes abrasion arthroplasty and drilling. Due to the limitations of the methodology of searching the database for this code, it is unclear as to how many patients underwent actual microfracture vs abrasion arthroplasty.

CONCLUSION

Within a large US commercial insurance database from 2007 to 2011, reparative procedures were favored for chondral injuries, but yielded an increased risk for conversion to arthroplasty. There was no difference between failure/revision rates among the restorative approaches, yet cell-based approaches yielded a significantly increased risk for a return to the OR.

Seventy-five percent of adults aged >75 years have hypertension.^1-3 According to the Joint National Commission 8 (JNC 8), the recommended target blood pressure (BP) is < 150/80 mm Hg for adults aged > 60 years.⁴ In 2016 the Systolic Blood Pressure Intervention Trial (SPRINT) suggested that more aggressive BP control with a goal of < 120/80 mm Hg reduced rates of cardiovascular disease and lowered the risk of death in adults aged > 50 years with hypertension.⁵ It is anticipated that as a result of the landmark SPRINT results, clinicians may attempt to treat hypertension more intensely in older patients with an increased risk of adverse consequences if BPs are not appropriately measured.

There is a standardized protocol for BP measurement, but these recommendations typically are not followed in routine office visits.^6,7 Some studies have noted that home BP measurement may be more accurate than office measurement.⁸ However, clinicians may not always trust the accuracy of home BP readings, and many patients are not adherent with home measurement. As a result, physicians usually manage hypertension in older patients based on office readings, though it is likely that most office measurements do not follow protocol on proper measurement. Office measurements have been noted to be inaccurate with high likelihood of overestimating or underestimating BP control.⁹

Office BP measurements demonstrate poor correlation with home measurements and have not been shown to be as good of a predictor for target organ damage or long-term cardiovascular outcomes compared with that of home measurements.^10,11 Although there have been studies comparing home and office BP measurements and comparing office and ambulatory BP measurement, no literature has been found that reports on the difference between routine office and standardized measurement of BP.^9,12-14

This study seeks to identify the magnitude of difference among BP measured according to a standardized protocol, routine clinical, and home BP. The authors hypothesized that there would be a significant, clinically relevant difference among the 3 BP measurement methods, especially between the routine office and standardized office measurements. This study has implications for implementing intensive treatment of hypertension based on office measurements.

Methods

Participants included 30 male veterans aged > 65 years who were actively participating in the Gerofit program at the VA Greater Los Angeles Healthcare System (VAGLAHS). The Gerofit program is a model clinical demonstration exercise and health promotion program targeting older and veterans at risk for falls or institutionalization. Gerofit was established in 1987 at the Durham VA Health System and successfully implemented in 2014 at VAGLAHS. Supervised exercise is offered 3 times per week and consists of individually tailored exercises aimed at reducing functional deficits that are identified and monitored by an initial and quarterly functional assessment. Blood pressures are checked routinely once a week as a part of the program. Gerofit was reviewed and approved by the institutional review board at VAGLAHS as a quality improvement/quality assurance project.

Data

Routine office and standardized protocol measurements were obtained by a single CasMED 740 (Branford, CT) automated BP machine and were conducted separately on different days. The CasMED 740 machine was not otherwise calibrated; however, a one-time correlation was performed between the CasMED 740 and the home BP monitor for each participant, when it was brought to VAGLAHS. Two measurements were made with the CasMed 740 automated BP machine on the arm that gave the higher BP reading throughout the standardized and routine protocol. Two subsequent measurements were made with the participant’s home automated BP cuff. Averages for the CasMED 740 and the home BP monitoring device were compared and assessed for significance by paired t test. No rest was scheduled prior to the first measurement, but there was a 1-minute rest after each subsequent measurement.

Mean values (SD) were used for participant characteristics and mean values (standard error [SE]) were used for BP measurements. Data were analyzed using Microsoft Excel (Redmond, WA) and GraphPad Prism version 7.03 (San Diego, CA). T tests were used for analysis of home BP measurements due to low sample size. Values of P < .05 were considered to be statistically significant.

Routine office protocol. Automated BP was measured to mimic routine office visits. Upon arrival, participants sat down, and the BP cuff was placed around their arm. Any rest before a measurement was incidental and not intentionally structured. Appropriate cuff size was determined by visual estimation of arm circumference. Only 1 measurement was made unless BP was > 150/90 mm Hg, in which case a repeat measurement was made after 2 to 4 minutes of rest. The BP was then determined based on the average of 2 or more readings. The BPs were recorded by hand in a weekly log. Participants had at least 12 weeks of BP readings measured by the routine method, and these BPs were averaged over 12 weeks to yield their average routine measured BP.

Standardized protocol. Automated BP was measured according to the 2015 USPSTF Guidelines and Look AHEAD trial protocol.^7,15 A participant’s arm circumference was measured, and appropriate cuff size was determined. The participant rested quietly in a chair for at least 5 minutes with feet flat on the floor and back supported. The cuff was snugly placed 2 to 3 cm above the antecubital fossa, and the arm was supported at the level of the right atrium during the measurement. Blood pressure was determined using the mean of 4 automated cuff readings, 2 on each arm, taken 1 minute apart. Participants did not necessarily have their BP measured by the standardized method immediately following the routine method but all measurements were performed during the same 12-week time period.

Home blood pressure protocol. Participants were given instructions according to the American Heart Association (AHA) recommendations for measuring home BP. Patients were instructed to use a calibrated, automated arm BP cuff. Home BP machines were not provided in advance, and each individual’s BP machine was not calibrated. They also were instructed to rest at least 5 minutes before measuring their BP. The mean home BP was determined by the cumulative average of 3 readings in the morning and evening, taken 1 minute between each reading, for a total of 6 readings/d. Participants recorded home BPs for 2 weeks before submitting their readings. Each participant affirmed clear understanding of how to measure BP by correctly demonstrating placement of the cuff 1 time under supervision.

 

Results

Thirty veterans aged > 65 years participated in the study. The average age (SD) was 82.7 (9.3) years. The average BMI (SD) was overweight at 29.7 kg/m² (5.7). Most (87.6%) of the study participants had been diagnosed with hypertension prior to the study, and no new diagnoses were made as a result of the study. 

Participants were prescribed an average of 1.3 antihypertensive medications and 0.8 medications that had BP effects (Table 1).

Both systolic BP (SBP) and diastolic BPs (DBP) measured by the standardized method were significantly lower than those by the routine method (P < .01 and P < .01, respectively) (Figure 1). 

The average SBP measured by routine method was 135 mm Hg compared with 115 mm Hg for the standardized method (Table 2). The average routine method DBP was 73 mm Hg vs 62 mm Hg by the standardized method (Table 2). Home BPs approximated routine BPs more closely with an average SBP of 129 mm Hg and DBP of 69 mm Hg. All participants were given instructions about how to monitor BP, but only 13 out of 30 returned completed home BP logs. There was no statistically significant difference between home and routine or between home and standardized BP readings.

To determine the accuracy of the home BP monitors, the average routine VAGLAHS BP measure was compared with home BP results. For SBPs, there was a significant correlation coefficient of 0.91 (P < .01). 

For DBPs, the correlation coefficient was 0.97 and were also significant (P < .01) (Figure 2).

Discussion

The present study demonstrated that standardized measurements of BP were lower than that of the routine method used in most office settings. These results suggest that there could be a risk of overtreatment for some patients those of whose results are higher than the SPRINT BP target of < 120/80 mm Hg. Clinicians might be treating BPs that are elevated due to improper measurement, which can lead to deleterious consequences in older adults, such as syncope and falls.¹⁶

Each participant exhibited a significantly lower BP reading with the standardized method than the routine method. The 20-point decrease in SBP and 10-point decrease in DBP are clinically significant. The routine method of measurement was intended to simulate BP measurement in outpatient settings. There is usually little time structured for rest, and because the protocol established by the AHA and other professional organizations is time consuming, it usually is not strictly followed. With guidelines proposed by JNC 8 and new findings from SPRINT, the method of BP readings should be reviewed in all clinical settings.

While changes in BP management are not necessarily immediate, the differences in recommendations proposed by SPRINT and JNC 8 can lead to confusion regarding how intensely to treat BP. These recommendations guide clinical practice, but clinicians’ best judgment ultimately determines BP management. Physicians who utilize routine office measurements likely rely on BP readings that are higher on average than are readings done under proper conditions. This leads to the prospect of overtreatment, where physicians attempt to control hypertension too aggressively, potentially leading to orthostatic hypotension, syncope, and increased risk for falls.¹⁶ With findings from SPRINT recommending even lower BPs than that by JNC 8, overtreatment risk becomes especially relevant. While BP protocol was strictly followed in SPRINT, some clinicians may not necessarily follow the same fastidious protocol.

The average differences between the home and standardized BPs were not statistically significant possibly due to the small sample size in the home BP measurements; however, the difference might represent some clinical relevance. There was a 15-point difference in SBP results between home (129 mm Hg) and standardized (115 mm Hg) measures. There also was a difference in DBP between home (69 mm Hg) and standardized (62 mm Hg) results. The close correlation between both home and BPs measured in VAGLAHS demonstrated that any difference was not due to variability in the measurement devices. Previous studies have demonstrated that home BPs are better indicators of cardiovascular risk than office BP.⁸

Despite lack of statistical significance, home BPs were lower than routine, which suggests that they still may be more reliable than routine office measurements. Definitive conclusions regarding the accuracy of the home BPs in the present study cannot be drawn due to the small sample size (n = 13). Further exploration with comparisons to ambulatory BP monitoring could yield more information on accuracy of home BP monitoring.

In this study’s cohort of older veterans, the average BMI was between 25 and 30 (overweight), which is a risk factor for hypertension.¹⁷ Every participant with hypertension was taking at least 1 antihypertensive medication and being actively managed. In this study, the authors accounted for other medications that may affect BP, such as α blockers used in patients with benign prostatic hyperplasia.¹⁸ These could have potential elevating or lowering effects on BP measurements.

An issue in this study was the lack of adherence to home BP monitoring. Many patients forgot to bring in their records or to measure their BPs at home. The difficulties highlight real-life issues. Clinicians often request that patients monitor their BP at home, but few may actually remember, let alone keep diligent records. There are many barriers between measuring and reporting home BPs, which may prevent the usefulness of monitoring BP at home.

 

Limitations

There were several limitations to the study. There was no specific protocol for the routine method of BP measurement, as it was intended to simulate the haphazard nature of office measurements. However, this approach limits its reproducibility. For home BP monitoring, it would have been ideal to provide the same calibrated, automated BP device to each participant. This study of older veterans may not be applicable to the general population. Finally, the relatively small number of participants in the study (n = 30) may have limited power in drawing definitive conclusions.

Future Directions

For future studies, comparing the standardized method to ambulatory BP monitoring would provide more information on accuracy. In addition, the authors would like to evaluate the effect of exercise on BP measurements in the different settings: home, standardized, and routine methods.

Seventy-five percent of adults aged >75 years have hypertension.^1-3 According to the Joint National Commission 8 (JNC 8), the recommended target blood pressure (BP) is < 150/80 mm Hg for adults aged > 60 years.⁴ In 2016 the Systolic Blood Pressure Intervention Trial (SPRINT) suggested that more aggressive BP control with a goal of < 120/80 mm Hg reduced rates of cardiovascular disease and lowered the risk of death in adults aged > 50 years with hypertension.⁵ It is anticipated that as a result of the landmark SPRINT results, clinicians may attempt to treat hypertension more intensely in older patients with an increased risk of adverse consequences if BPs are not appropriately measured.

There is a standardized protocol for BP measurement, but these recommendations typically are not followed in routine office visits.^6,7 Some studies have noted that home BP measurement may be more accurate than office measurement.⁸ However, clinicians may not always trust the accuracy of home BP readings, and many patients are not adherent with home measurement. As a result, physicians usually manage hypertension in older patients based on office readings, though it is likely that most office measurements do not follow protocol on proper measurement. Office measurements have been noted to be inaccurate with high likelihood of overestimating or underestimating BP control.⁹

Office BP measurements demonstrate poor correlation with home measurements and have not been shown to be as good of a predictor for target organ damage or long-term cardiovascular outcomes compared with that of home measurements.^10,11 Although there have been studies comparing home and office BP measurements and comparing office and ambulatory BP measurement, no literature has been found that reports on the difference between routine office and standardized measurement of BP.^9,12-14

This study seeks to identify the magnitude of difference among BP measured according to a standardized protocol, routine clinical, and home BP. The authors hypothesized that there would be a significant, clinically relevant difference among the 3 BP measurement methods, especially between the routine office and standardized office measurements. This study has implications for implementing intensive treatment of hypertension based on office measurements.

Methods

Participants included 30 male veterans aged > 65 years who were actively participating in the Gerofit program at the VA Greater Los Angeles Healthcare System (VAGLAHS). The Gerofit program is a model clinical demonstration exercise and health promotion program targeting older and veterans at risk for falls or institutionalization. Gerofit was established in 1987 at the Durham VA Health System and successfully implemented in 2014 at VAGLAHS. Supervised exercise is offered 3 times per week and consists of individually tailored exercises aimed at reducing functional deficits that are identified and monitored by an initial and quarterly functional assessment. Blood pressures are checked routinely once a week as a part of the program. Gerofit was reviewed and approved by the institutional review board at VAGLAHS as a quality improvement/quality assurance project.

Data

Routine office and standardized protocol measurements were obtained by a single CasMED 740 (Branford, CT) automated BP machine and were conducted separately on different days. The CasMED 740 machine was not otherwise calibrated; however, a one-time correlation was performed between the CasMED 740 and the home BP monitor for each participant, when it was brought to VAGLAHS. Two measurements were made with the CasMed 740 automated BP machine on the arm that gave the higher BP reading throughout the standardized and routine protocol. Two subsequent measurements were made with the participant’s home automated BP cuff. Averages for the CasMED 740 and the home BP monitoring device were compared and assessed for significance by paired t test. No rest was scheduled prior to the first measurement, but there was a 1-minute rest after each subsequent measurement.

Mean values (SD) were used for participant characteristics and mean values (standard error [SE]) were used for BP measurements. Data were analyzed using Microsoft Excel (Redmond, WA) and GraphPad Prism version 7.03 (San Diego, CA). T tests were used for analysis of home BP measurements due to low sample size. Values of P < .05 were considered to be statistically significant.

Routine office protocol. Automated BP was measured to mimic routine office visits. Upon arrival, participants sat down, and the BP cuff was placed around their arm. Any rest before a measurement was incidental and not intentionally structured. Appropriate cuff size was determined by visual estimation of arm circumference. Only 1 measurement was made unless BP was > 150/90 mm Hg, in which case a repeat measurement was made after 2 to 4 minutes of rest. The BP was then determined based on the average of 2 or more readings. The BPs were recorded by hand in a weekly log. Participants had at least 12 weeks of BP readings measured by the routine method, and these BPs were averaged over 12 weeks to yield their average routine measured BP.

Standardized protocol. Automated BP was measured according to the 2015 USPSTF Guidelines and Look AHEAD trial protocol.^7,15 A participant’s arm circumference was measured, and appropriate cuff size was determined. The participant rested quietly in a chair for at least 5 minutes with feet flat on the floor and back supported. The cuff was snugly placed 2 to 3 cm above the antecubital fossa, and the arm was supported at the level of the right atrium during the measurement. Blood pressure was determined using the mean of 4 automated cuff readings, 2 on each arm, taken 1 minute apart. Participants did not necessarily have their BP measured by the standardized method immediately following the routine method but all measurements were performed during the same 12-week time period.

Home blood pressure protocol. Participants were given instructions according to the American Heart Association (AHA) recommendations for measuring home BP. Patients were instructed to use a calibrated, automated arm BP cuff. Home BP machines were not provided in advance, and each individual’s BP machine was not calibrated. They also were instructed to rest at least 5 minutes before measuring their BP. The mean home BP was determined by the cumulative average of 3 readings in the morning and evening, taken 1 minute between each reading, for a total of 6 readings/d. Participants recorded home BPs for 2 weeks before submitting their readings. Each participant affirmed clear understanding of how to measure BP by correctly demonstrating placement of the cuff 1 time under supervision.

 

Results

Thirty veterans aged > 65 years participated in the study. The average age (SD) was 82.7 (9.3) years. The average BMI (SD) was overweight at 29.7 kg/m² (5.7). Most (87.6%) of the study participants had been diagnosed with hypertension prior to the study, and no new diagnoses were made as a result of the study. 

Participants were prescribed an average of 1.3 antihypertensive medications and 0.8 medications that had BP effects (Table 1).

Both systolic BP (SBP) and diastolic BPs (DBP) measured by the standardized method were significantly lower than those by the routine method (P < .01 and P < .01, respectively) (Figure 1). 

The average SBP measured by routine method was 135 mm Hg compared with 115 mm Hg for the standardized method (Table 2). The average routine method DBP was 73 mm Hg vs 62 mm Hg by the standardized method (Table 2). Home BPs approximated routine BPs more closely with an average SBP of 129 mm Hg and DBP of 69 mm Hg. All participants were given instructions about how to monitor BP, but only 13 out of 30 returned completed home BP logs. There was no statistically significant difference between home and routine or between home and standardized BP readings.

To determine the accuracy of the home BP monitors, the average routine VAGLAHS BP measure was compared with home BP results. For SBPs, there was a significant correlation coefficient of 0.91 (P < .01). 

For DBPs, the correlation coefficient was 0.97 and were also significant (P < .01) (Figure 2).

Discussion

The present study demonstrated that standardized measurements of BP were lower than that of the routine method used in most office settings. These results suggest that there could be a risk of overtreatment for some patients those of whose results are higher than the SPRINT BP target of < 120/80 mm Hg. Clinicians might be treating BPs that are elevated due to improper measurement, which can lead to deleterious consequences in older adults, such as syncope and falls.¹⁶

Each participant exhibited a significantly lower BP reading with the standardized method than the routine method. The 20-point decrease in SBP and 10-point decrease in DBP are clinically significant. The routine method of measurement was intended to simulate BP measurement in outpatient settings. There is usually little time structured for rest, and because the protocol established by the AHA and other professional organizations is time consuming, it usually is not strictly followed. With guidelines proposed by JNC 8 and new findings from SPRINT, the method of BP readings should be reviewed in all clinical settings.

While changes in BP management are not necessarily immediate, the differences in recommendations proposed by SPRINT and JNC 8 can lead to confusion regarding how intensely to treat BP. These recommendations guide clinical practice, but clinicians’ best judgment ultimately determines BP management. Physicians who utilize routine office measurements likely rely on BP readings that are higher on average than are readings done under proper conditions. This leads to the prospect of overtreatment, where physicians attempt to control hypertension too aggressively, potentially leading to orthostatic hypotension, syncope, and increased risk for falls.¹⁶ With findings from SPRINT recommending even lower BPs than that by JNC 8, overtreatment risk becomes especially relevant. While BP protocol was strictly followed in SPRINT, some clinicians may not necessarily follow the same fastidious protocol.

The average differences between the home and standardized BPs were not statistically significant possibly due to the small sample size in the home BP measurements; however, the difference might represent some clinical relevance. There was a 15-point difference in SBP results between home (129 mm Hg) and standardized (115 mm Hg) measures. There also was a difference in DBP between home (69 mm Hg) and standardized (62 mm Hg) results. The close correlation between both home and BPs measured in VAGLAHS demonstrated that any difference was not due to variability in the measurement devices. Previous studies have demonstrated that home BPs are better indicators of cardiovascular risk than office BP.⁸

Despite lack of statistical significance, home BPs were lower than routine, which suggests that they still may be more reliable than routine office measurements. Definitive conclusions regarding the accuracy of the home BPs in the present study cannot be drawn due to the small sample size (n = 13). Further exploration with comparisons to ambulatory BP monitoring could yield more information on accuracy of home BP monitoring.

In this study’s cohort of older veterans, the average BMI was between 25 and 30 (overweight), which is a risk factor for hypertension.¹⁷ Every participant with hypertension was taking at least 1 antihypertensive medication and being actively managed. In this study, the authors accounted for other medications that may affect BP, such as α blockers used in patients with benign prostatic hyperplasia.¹⁸ These could have potential elevating or lowering effects on BP measurements.

An issue in this study was the lack of adherence to home BP monitoring. Many patients forgot to bring in their records or to measure their BPs at home. The difficulties highlight real-life issues. Clinicians often request that patients monitor their BP at home, but few may actually remember, let alone keep diligent records. There are many barriers between measuring and reporting home BPs, which may prevent the usefulness of monitoring BP at home.

 

Limitations

There were several limitations to the study. There was no specific protocol for the routine method of BP measurement, as it was intended to simulate the haphazard nature of office measurements. However, this approach limits its reproducibility. For home BP monitoring, it would have been ideal to provide the same calibrated, automated BP device to each participant. This study of older veterans may not be applicable to the general population. Finally, the relatively small number of participants in the study (n = 30) may have limited power in drawing definitive conclusions.

Future Directions

For future studies, comparing the standardized method to ambulatory BP monitoring would provide more information on accuracy. In addition, the authors would like to evaluate the effect of exercise on BP measurements in the different settings: home, standardized, and routine methods.

Seventy-five percent of adults aged >75 years have hypertension.^1-3 According to the Joint National Commission 8 (JNC 8), the recommended target blood pressure (BP) is < 150/80 mm Hg for adults aged > 60 years.⁴ In 2016 the Systolic Blood Pressure Intervention Trial (SPRINT) suggested that more aggressive BP control with a goal of < 120/80 mm Hg reduced rates of cardiovascular disease and lowered the risk of death in adults aged > 50 years with hypertension.⁵ It is anticipated that as a result of the landmark SPRINT results, clinicians may attempt to treat hypertension more intensely in older patients with an increased risk of adverse consequences if BPs are not appropriately measured.

There is a standardized protocol for BP measurement, but these recommendations typically are not followed in routine office visits.^6,7 Some studies have noted that home BP measurement may be more accurate than office measurement.⁸ However, clinicians may not always trust the accuracy of home BP readings, and many patients are not adherent with home measurement. As a result, physicians usually manage hypertension in older patients based on office readings, though it is likely that most office measurements do not follow protocol on proper measurement. Office measurements have been noted to be inaccurate with high likelihood of overestimating or underestimating BP control.⁹

Office BP measurements demonstrate poor correlation with home measurements and have not been shown to be as good of a predictor for target organ damage or long-term cardiovascular outcomes compared with that of home measurements.^10,11 Although there have been studies comparing home and office BP measurements and comparing office and ambulatory BP measurement, no literature has been found that reports on the difference between routine office and standardized measurement of BP.^9,12-14

This study seeks to identify the magnitude of difference among BP measured according to a standardized protocol, routine clinical, and home BP. The authors hypothesized that there would be a significant, clinically relevant difference among the 3 BP measurement methods, especially between the routine office and standardized office measurements. This study has implications for implementing intensive treatment of hypertension based on office measurements.

Methods

Participants included 30 male veterans aged > 65 years who were actively participating in the Gerofit program at the VA Greater Los Angeles Healthcare System (VAGLAHS). The Gerofit program is a model clinical demonstration exercise and health promotion program targeting older and veterans at risk for falls or institutionalization. Gerofit was established in 1987 at the Durham VA Health System and successfully implemented in 2014 at VAGLAHS. Supervised exercise is offered 3 times per week and consists of individually tailored exercises aimed at reducing functional deficits that are identified and monitored by an initial and quarterly functional assessment. Blood pressures are checked routinely once a week as a part of the program. Gerofit was reviewed and approved by the institutional review board at VAGLAHS as a quality improvement/quality assurance project.

Data

Routine office and standardized protocol measurements were obtained by a single CasMED 740 (Branford, CT) automated BP machine and were conducted separately on different days. The CasMED 740 machine was not otherwise calibrated; however, a one-time correlation was performed between the CasMED 740 and the home BP monitor for each participant, when it was brought to VAGLAHS. Two measurements were made with the CasMed 740 automated BP machine on the arm that gave the higher BP reading throughout the standardized and routine protocol. Two subsequent measurements were made with the participant’s home automated BP cuff. Averages for the CasMED 740 and the home BP monitoring device were compared and assessed for significance by paired t test. No rest was scheduled prior to the first measurement, but there was a 1-minute rest after each subsequent measurement.

Mean values (SD) were used for participant characteristics and mean values (standard error [SE]) were used for BP measurements. Data were analyzed using Microsoft Excel (Redmond, WA) and GraphPad Prism version 7.03 (San Diego, CA). T tests were used for analysis of home BP measurements due to low sample size. Values of P < .05 were considered to be statistically significant.

Routine office protocol. Automated BP was measured to mimic routine office visits. Upon arrival, participants sat down, and the BP cuff was placed around their arm. Any rest before a measurement was incidental and not intentionally structured. Appropriate cuff size was determined by visual estimation of arm circumference. Only 1 measurement was made unless BP was > 150/90 mm Hg, in which case a repeat measurement was made after 2 to 4 minutes of rest. The BP was then determined based on the average of 2 or more readings. The BPs were recorded by hand in a weekly log. Participants had at least 12 weeks of BP readings measured by the routine method, and these BPs were averaged over 12 weeks to yield their average routine measured BP.

Standardized protocol. Automated BP was measured according to the 2015 USPSTF Guidelines and Look AHEAD trial protocol.^7,15 A participant’s arm circumference was measured, and appropriate cuff size was determined. The participant rested quietly in a chair for at least 5 minutes with feet flat on the floor and back supported. The cuff was snugly placed 2 to 3 cm above the antecubital fossa, and the arm was supported at the level of the right atrium during the measurement. Blood pressure was determined using the mean of 4 automated cuff readings, 2 on each arm, taken 1 minute apart. Participants did not necessarily have their BP measured by the standardized method immediately following the routine method but all measurements were performed during the same 12-week time period.

Home blood pressure protocol. Participants were given instructions according to the American Heart Association (AHA) recommendations for measuring home BP. Patients were instructed to use a calibrated, automated arm BP cuff. Home BP machines were not provided in advance, and each individual’s BP machine was not calibrated. They also were instructed to rest at least 5 minutes before measuring their BP. The mean home BP was determined by the cumulative average of 3 readings in the morning and evening, taken 1 minute between each reading, for a total of 6 readings/d. Participants recorded home BPs for 2 weeks before submitting their readings. Each participant affirmed clear understanding of how to measure BP by correctly demonstrating placement of the cuff 1 time under supervision.

 

Results

Thirty veterans aged > 65 years participated in the study. The average age (SD) was 82.7 (9.3) years. The average BMI (SD) was overweight at 29.7 kg/m² (5.7). Most (87.6%) of the study participants had been diagnosed with hypertension prior to the study, and no new diagnoses were made as a result of the study. 

Participants were prescribed an average of 1.3 antihypertensive medications and 0.8 medications that had BP effects (Table 1).

Both systolic BP (SBP) and diastolic BPs (DBP) measured by the standardized method were significantly lower than those by the routine method (P < .01 and P < .01, respectively) (Figure 1). 

The average SBP measured by routine method was 135 mm Hg compared with 115 mm Hg for the standardized method (Table 2). The average routine method DBP was 73 mm Hg vs 62 mm Hg by the standardized method (Table 2). Home BPs approximated routine BPs more closely with an average SBP of 129 mm Hg and DBP of 69 mm Hg. All participants were given instructions about how to monitor BP, but only 13 out of 30 returned completed home BP logs. There was no statistically significant difference between home and routine or between home and standardized BP readings.

To determine the accuracy of the home BP monitors, the average routine VAGLAHS BP measure was compared with home BP results. For SBPs, there was a significant correlation coefficient of 0.91 (P < .01). 

For DBPs, the correlation coefficient was 0.97 and were also significant (P < .01) (Figure 2).

Discussion

The present study demonstrated that standardized measurements of BP were lower than that of the routine method used in most office settings. These results suggest that there could be a risk of overtreatment for some patients those of whose results are higher than the SPRINT BP target of < 120/80 mm Hg. Clinicians might be treating BPs that are elevated due to improper measurement, which can lead to deleterious consequences in older adults, such as syncope and falls.¹⁶

Each participant exhibited a significantly lower BP reading with the standardized method than the routine method. The 20-point decrease in SBP and 10-point decrease in DBP are clinically significant. The routine method of measurement was intended to simulate BP measurement in outpatient settings. There is usually little time structured for rest, and because the protocol established by the AHA and other professional organizations is time consuming, it usually is not strictly followed. With guidelines proposed by JNC 8 and new findings from SPRINT, the method of BP readings should be reviewed in all clinical settings.

While changes in BP management are not necessarily immediate, the differences in recommendations proposed by SPRINT and JNC 8 can lead to confusion regarding how intensely to treat BP. These recommendations guide clinical practice, but clinicians’ best judgment ultimately determines BP management. Physicians who utilize routine office measurements likely rely on BP readings that are higher on average than are readings done under proper conditions. This leads to the prospect of overtreatment, where physicians attempt to control hypertension too aggressively, potentially leading to orthostatic hypotension, syncope, and increased risk for falls.¹⁶ With findings from SPRINT recommending even lower BPs than that by JNC 8, overtreatment risk becomes especially relevant. While BP protocol was strictly followed in SPRINT, some clinicians may not necessarily follow the same fastidious protocol.

The average differences between the home and standardized BPs were not statistically significant possibly due to the small sample size in the home BP measurements; however, the difference might represent some clinical relevance. There was a 15-point difference in SBP results between home (129 mm Hg) and standardized (115 mm Hg) measures. There also was a difference in DBP between home (69 mm Hg) and standardized (62 mm Hg) results. The close correlation between both home and BPs measured in VAGLAHS demonstrated that any difference was not due to variability in the measurement devices. Previous studies have demonstrated that home BPs are better indicators of cardiovascular risk than office BP.⁸

Despite lack of statistical significance, home BPs were lower than routine, which suggests that they still may be more reliable than routine office measurements. Definitive conclusions regarding the accuracy of the home BPs in the present study cannot be drawn due to the small sample size (n = 13). Further exploration with comparisons to ambulatory BP monitoring could yield more information on accuracy of home BP monitoring.

In this study’s cohort of older veterans, the average BMI was between 25 and 30 (overweight), which is a risk factor for hypertension.¹⁷ Every participant with hypertension was taking at least 1 antihypertensive medication and being actively managed. In this study, the authors accounted for other medications that may affect BP, such as α blockers used in patients with benign prostatic hyperplasia.¹⁸ These could have potential elevating or lowering effects on BP measurements.

An issue in this study was the lack of adherence to home BP monitoring. Many patients forgot to bring in their records or to measure their BPs at home. The difficulties highlight real-life issues. Clinicians often request that patients monitor their BP at home, but few may actually remember, let alone keep diligent records. There are many barriers between measuring and reporting home BPs, which may prevent the usefulness of monitoring BP at home.

 

Limitations

There were several limitations to the study. There was no specific protocol for the routine method of BP measurement, as it was intended to simulate the haphazard nature of office measurements. However, this approach limits its reproducibility. For home BP monitoring, it would have been ideal to provide the same calibrated, automated BP device to each participant. This study of older veterans may not be applicable to the general population. Finally, the relatively small number of participants in the study (n = 30) may have limited power in drawing definitive conclusions.

Future Directions

For future studies, comparing the standardized method to ambulatory BP monitoring would provide more information on accuracy. In addition, the authors would like to evaluate the effect of exercise on BP measurements in the different settings: home, standardized, and routine methods.

ABSTRACT

Use of mesenchymal stem cells from bone marrow has gained significant popularity. The iliac crest has been determined to be an effective site for harvesting mesenchymal stem cells. Review of the literature reveals that multiple techniques are used to harvest bone marrow aspirate from the iliac crest, but the descriptions are based on the experience of various authors as opposed to studied anatomy. A safe, reliable, and reproducible method for aspiration has yet to be studied and described. We hypothesized that there would be an ideal angle and distance for aspiration that would be the safest, most consistent, and most reliable. Using magnetic resonance imaging (MRI), we reviewed 26 total lumbar spine MRI scans (13 males, 13 females) and found that an angle of 24° should be used when entering the most medial aspect of the posterior superior iliac spine (PSIS) and that this angle did not differ between the sexes. The distance that the trocar can advance after entry before hitting the anterior ilium wall varied significantly between males and females, being 7.53 cm in males and 6.74 cm in females. In addition, the size of the PSIS table was significantly different between males and females (1.20 cm and 0.96 cm, respectively). No other significant differences in the measurements gathered were found. Using the data gleaned from this study, we developed an aspiration technique. This method uses ultrasound to determine the location of the PSIS and the entry point on the PSIS. This contrasts with most techniques that use landmark palpation, which is known to be unreliable and inaccurate. The described technique for aspiration from the PSIS is safe, reliable, reproducible, and substantiated by data.

The iliac crest is an effective site for harvesting bone marrow stem cells. It allows for easy access and is superficial in most individuals, allowing for a relatively quick and simple procedure. Use of mesenchymal stem cells (MSCs) for treatment of orthopedic injuries has grown recently. Whereas overall use has increased, review of the literature reveals very few techniques for iliac crest aspiration,¹ but these are not based on anatomic relationships or studies. Hernigou and colleagues^2,3 attempted to quantitatively evaluate potential “sectors” allowing for safe aspiration using cadaver and computed tomographic reconstruction imaging. We used magnetic resonance imaging (MRI) to analyze aspiration parameters. Owing to the ilium’s anatomy, improper positioning or aspiration technique during aspiration can result in serious injury.^2,4-6 We hypothesized that there is an ideal angle and positioning for bone marrow aspiration from the posterior superior iliac spine (PSIS) that is safe, consistent, and reproducible. Although most aspiration techniques use landmark palpation, this is unreliable and inaccurate, especially when compared with ultrasound-guided injections^7-16 and procedures.^9,12,17-19 We describe our technique using ultrasound to visualize patient anatomy and accurately determine anatomic entry with the trocar.

METHODS

MRI scans of 26 patients (13 males, 13 females) were reviewed to determine average angles and distances. Axial T2-weighted views of the lumbar spine were used in all analyses. The sacroiliac (SI) joint angle was defined as the angle formed between the vector through the midline of the pelvis and the vector that is parallel to the SI joint. The approach angle was defined as the angle formed between the vector of the most medial aspect of the PSIS through the ilium to the anterior wall and the vector through the midline of the pelvis (Figure 1). 

The distance, in centimeters, of the PSIS to the anterior ilium wall was measured to determine the maximum distance the trocar can advance without puncturing the anterior ilium wall. The perpendicular distance from the PSIS table to the anterior aspect of the ilium was measured to determine the average depth the trocar could advance if the clinician entered it perpendicular to the patient’s coronal plane (Figure 2). 

The PSIS table width was defined as the width, in centimeters, of the PSIS posteriorly. The minimum width, in centimeters, of the ilium was measured at the thinnest region of the ilium to determine the degree of variation if the trocar were entered too far laterally and advanced to the anterior ilium wall (Figure 3). 

Means and standard deviations were calculated by sex and as a total. Student’s t test was used to determine statistical significance (P < .05).

Continue to: For the 13 males, the mean SI joint...

 

RESULTS

The results are reported in the Table. 

Table. Measurements of Patients Taken on Axial T2-Weighted Views of Lumbosacral MRI Scans^a

Patient	SI Joint Angle (°)	Approach Angle (°)	PSIS Table Width (cm)	PSIS to Anterior Ilium Wall (cm)	Perpendicular Distance PSIS to Anterior Joint (cm)	Post Ilium Wall to SI Joint Width (cm)
Males
1	28.80	19.50	1.24	8.80	4.16	1.52
2	31.80	27.60	1.70	7.89	3.49	1.02
3	33.70	27.70	1.12	8.14	3.15	1.28
4	23.70	26.40	0.95	6.66	3.22	0.65
5	35.90	28.40	0.84	7.60	2.57	0.95
6	33.80	29.30	1.20	7.73	2.34	0.90
7	30.30	21.20	1.36	8.44	3.95	1.18
8	34.50	20.40	1.53	7.08	3.98	1.56
9	28.70	24.00	1.34	8.19	3.51	1.31
10	22.40	20.10	1.37	7.30	3.87	1.28
11	33.60	20.80	0.88	6.43	3.26	0.94
12	48.50	31.00	1.15	6.69	2.97	1.38
13	20.20	20.90	0.94	6.95	3.79	1.05
Averages	31.22	24.41	1.20	7.53	3.40	1.16
Standard Deviation	7.18	4.11	0.26	0.75	0.56	0.26
Females
14	22.80	23.20	1.54	7.21	3.45	1.39
15	33.30	21.40	1.09	7.26	3.57	0.98
16	19.70	15.60	0.78	8.32	3.76	0.86
17	17.50	15.60	0.61	7.57	3.37	1.03
18	48.20	26.60	0.94	6.62	3.16	0.71
19	38.20	28.30	0.90	6.32	2.23	0.91
20	44.50	31.70	0.99	6.19	3.06	0.76
21	24.10	18.00	0.92	6.99	3.23	0.71
22	17.20	14.80	0.81	6.00	2.81	1.13
23	42.00	38.50	1.00	5.33	2.47	1.42
24	32.00	25.50	0.98	6.01	2.79	1.21
25	24.70	24.80	0.87	6.09	2.79	1.02
26	19.80	22.30	1.04	7.71	2.37	1.36
Averages	29.54	23.56	0.96	6.74	3.00	1.04
Standard Deviation	10.84	6.88	0.21	0.85	0.48	0.25
All patients Averages	30.38	23.98	1.08	7.14	3.20	1.10
Standard Deviation	9.05	5.57	0.26	0.88	0.55	0.26

^aStatistical significance is denoted as P < .02.

Abbreviations: MRI, magnetic resonance imaging; PSIS, posterior iliac spine; SI, sacroiliac.

For the 13 males, the mean SI joint angle was 31.22° ± 7.18° (range, 20.20° to 48.50°). The mean approach angle was 24.41° ± 4.11° (range, 19.50° to 31.00°). The mean PSIS table width was 1.20 cm ± 0.26 cm (range, 0.84 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.53 cm ± 0.75 cm (range, 6.43 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.40 cm ± 0.56 cm (range, 2.34 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.16 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

For the 13 females, the mean SI joint angle was 29.54° ± 10.84° (range, 17.20° to 48.20°). The mean approach angle was 23.56° ± 6.88° (range, 14.80° to 38.50°). The mean PSIS table width was 0.96 cm ± 0.21 cm (range, 0.61 cm to 1.54 cm). The mean distance from the PSIS to the anterior ilium wall was 6.74 cm ± 0.85 cm (range, 5.33 cm to 8.32 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.00 cm ± 0.48 cm (range, 2.23 cm to 3.76 cm). The mean minimum width of the ilium to the SI joint was 1.04 cm ± 0.25 cm (range, 0.71 cm to 1.42 cm).

For the 26 total patients, the mean SI joint angle was 30.38° ± 9.05° (range, 17.20° to 48.50°). The mean approach angle was 23.98° ± 5.57° (range, 14.80° to 38.50°). The mean PSIS table width was 1.08 cm ± 0.26 cm (range, 0.61 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.14 cm ± 0.88 cm (range, 5.33 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.20 cm ± 0.55 cm (range, 2.23 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.10 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

There was a statistically significant difference between the male and female groups for the maximum distance the trocar can be advanced from the PSIS to the anterior ilium wall (P < .02), and a statistically significant difference for the PSIS table width (P < .02). There were no significant differences between the male and female groups for the approach angle, the SI joint angle, the perpendicular distance from the PSIS to the anterior ilium, and the minimum width of the ilium to the SI joint.

Continue to: The patient is brought to the procedure...

 

TECHNIQUE: ILIAC CREST (PSIS) BONE MARROW ASPIRATION

The patient is brought to the procedure room and placed in a prone position. The donor site is prepared and draped in the usual sterile manner. Ultrasound is used to identify the median sacral crest in a short-axis view. The probe is then moved laterally to identify the PSIS (Figures 4A, 4B). 

The probe can be moved superiorly and inferiorly to determine the most prominent and central portion of the PSIS. The SI joint and ilium can also be visualized if needed.

The crosshairs on the ultrasound probe are used to mark the center lines of each plane. The central point marks the location of the PSIS. Alternatively, an in-plane technique can be used to place a spinal needle on the exact entry point on the PSIS. Once the PSIS and entry point are identified, the site is blocked with 10 mL of 0.5% ropivacaine.

Prior to introduction of the trocar, all instrumentation is primed with heparin and syringes are prepped with anticoagulant citrate dextrose solution, solution A. A stab incision is made at the site. The trocar is placed at the entry point, which should be centered in a superior-inferior plane and at the most medial point of the PSIS. Starting with the trocar vertical, the trocar is angled laterally 24° by dropping the hand medially toward the midline. No angulation cephalad or caudad is necessary, but cephalad must be avoided so as not to skive superiorly. This angle, which is recommended for both males and females, allows for the greatest distance the trocar can travel in bone before hitting the anterior ilium wall. A standard deviation of 5.57° is present, which should be considered. Steady pressure should be applied with a slight twisting motion on the PSIS. If advancement of the trocar is too difficult, a mallet or drill can be used to assist in penetration.

With the trocar advanced into the bone 1 cm, the trocar needle is removed while the cannula remains in place. The syringe is attached to the top of the cannula. The syringe plunger is pulled back to aspirate 20 mL of bone marrow. The cannula and syringe assembly are advanced 2 cm farther into the bone to allow for aspiration of a new location within the bone marrow cavity, and 20 mL of bone marrow are again aspirated. This is done a final time, advancing the trocar another 2 cm and aspirating a final 20 mL of bone marrow. The entire process should yield roughly 60 mL of bone marrow from one side. If desired, the same process can be repeated for the contralateral PSIS to yield a total of 120 mL of bone marrow from the 2 sites.

Based on our data, the average distance to the anterior ilium wall was 7 cm, but the shortest distance noted in this study was 5 cm. On the basis of the data presented, this technique allows for safe advancement based on even the shortest measured distance, without fear of puncturing the anterior ilium wall. Perforation could damage the femoral nerve and the internal or external iliac artery or vein that lie anterior to the ilium.

Continue to: We hypothesized that there...

 

DISCUSSION

We hypothesized that there would be an optimal angle of entry and maximal safe distance the trocar could advance through the ilium when aspirating. Because male and female pelvic anatomy differs, we also hypothesized that there would be differences in distance and size measurements for males and females. Our results supported our hypothesis that there is an ideal approach angle. The results also showed that the maximum distance the trocar can advance and the width of the PSIS table differ significantly between males and females.

Although pelvic anatomy differs between males and females, there should be an ideal entry angle that would allow maximum advancement into the ilium without perforating the anterior wall, which we defined as the approach angle. In our comparison of 26 MRI scans, we found that the approach angle did not differ significantly between the 2 groups (13 males, 13 females). This allows clinicians to enter the PSIS at roughly 24° medial to the parasagittal line, maximizing the space before puncturing into the anterior pelvis in either males or females.

If clinicians were to enter perpendicular to the patient’s PSIS, they would, on average, be able to advance only 3.20 cm before encountering the SI joint. When entering at 24° as we recommend, the average distance increases to 7.14 cm. Although the angle did not differ significantly, there was a significant difference between males and females in the length from the PSIS to the anterior wall, with males having 7.53 cm distance and females 6.74 cm. This is an important measurement because if the anterior ilium wall is punctured, the femoral nerve and the common, internal and external iliac arteries and veins could be damaged, resulting in retroperitoneal hemorrhage.

A fatality in 2001 in the United Kingdom led to a national audit of bone marrow aspiration and biopsies.^4-6 Although these procedures were done primarily for patients with cancer, hemorrhagic events were the most frequent and serious events. This audit led to the identification of many risk factors. Bain^4-6 conducted reviews of bone marrow aspirations and biopsies in the United Kingdom from 2002 to 2004. Of a total of 53,088 procedures conducted during that time frame, 48 (0.09%) adverse events occurred, with 29 (0.05%) being hemorrhagic events. Although infrequent, hemorrhagic adverse events represent significant morbidity. Reviews such as those conducted by Bain^4-6 highlight the importance of a study that helps determine the optimal parameters for aspiration to ensure safety and reliability.

Hernigou and colleagues^2,3 conducted studies analyzing different “sectors” in an attempt to develop a safe aspiration technique. They found that obese patients were at higher risk, and some sites of aspiration (sectors 1, 4, 5) had increased risk for perforation and damage to surrounding structures. Their sector 6, which incorporated the entirety of the PSIS table, was considered the safest, most reliable site for trocar introduction.^2,3 Hernigou and colleagues,² in comparing the bone mass of the sectors, also noted that sector 6 has the greatest bone thickness close to the entry point, making it the most favorable site. The PSIS is not just a point; it is more a “table.” The PSIS can be palpated posteriorly, but this is inaccurate and unreliable, particularly in larger individuals. The PSIS table can be identified on ultrasound before introducing the trocar, which is a more reliable method of landmark identification than palpation guidance, just as in ultrasound-guided injections^7-16 and procedures.^9,12,17-19

Continue to: If the PSIS is not accurately...

 

If the PSIS is not accurately identified, penetration laterally will result in entering the ilium wing, where it is quite narrow. We found the distance between the posterior ilium wall and the SI joint to be only 1.10 cm wide (Figure 3); we defined this area as the narrow corridor. Superior and lateral entry could damage the superior cluneal nerves coming over the iliac crest, which are located 6 cm lateral to the SI joint. Inferior and lateral entry 6 cm below the PSIS could reach the greater sciatic foramen, damaging the sacral plexus and superior gluteal artery and vein. If the entry slips above the PSIS over the pelvis, the trocar could enter the retroperitoneal space and damage the femoral nerve and common iliac artery and vein, leading to a retroperitoneal hemorrhage.^4-6,20

MSCs are found as perivascular cells and lie in the cortices of bones.²¹ Following the approach angle and directed line from the PSIS to the anterior ilium wall described in this study (Figures 1 and 2), the trocar would pass through the narrow corridor as it advances farther into the ilium. The minimum width of this corridor was measured in this study and, on average, was 1.10 cm wide from cortex to cortex (Figure 3). As the bone marrow is aspirated from this narrow corridor, the clinician is gathering MSCs from both the lateral and medial cortices of the ilium. By aspirating from a greater surface area of the cortices, it is believed that this will increase the total collection of MSCs.

CONCLUSION

Although there are reports in the literature that describe techniques for bone marrow aspiration from the iliac crest, the techniques are very general and vague regarding the ideal angles and methods. Studies have attempted to quantify the safest entry sites for aspiration but have not detailed ideal parameters for collection. Blind aspiration from the iliac crest can have serious implications if adverse events occur, and thus there is a need for a safe and reliable method of aspiration from the iliac crest. Ultrasound guidance to identify anatomy, as opposed to palpation guidance, ensures anatomic placement of the trocar while minimizing the risk of aspiration. Based on the measurements gathered in this study, an optimal angle of entry and safe distance of penetration have been identified. Using our data and relevant literature, we developed a technique for a safe, consistent, and reliable method of bone marrow aspiration out of the iliac crest.

ABSTRACT

Use of mesenchymal stem cells from bone marrow has gained significant popularity. The iliac crest has been determined to be an effective site for harvesting mesenchymal stem cells. Review of the literature reveals that multiple techniques are used to harvest bone marrow aspirate from the iliac crest, but the descriptions are based on the experience of various authors as opposed to studied anatomy. A safe, reliable, and reproducible method for aspiration has yet to be studied and described. We hypothesized that there would be an ideal angle and distance for aspiration that would be the safest, most consistent, and most reliable. Using magnetic resonance imaging (MRI), we reviewed 26 total lumbar spine MRI scans (13 males, 13 females) and found that an angle of 24° should be used when entering the most medial aspect of the posterior superior iliac spine (PSIS) and that this angle did not differ between the sexes. The distance that the trocar can advance after entry before hitting the anterior ilium wall varied significantly between males and females, being 7.53 cm in males and 6.74 cm in females. In addition, the size of the PSIS table was significantly different between males and females (1.20 cm and 0.96 cm, respectively). No other significant differences in the measurements gathered were found. Using the data gleaned from this study, we developed an aspiration technique. This method uses ultrasound to determine the location of the PSIS and the entry point on the PSIS. This contrasts with most techniques that use landmark palpation, which is known to be unreliable and inaccurate. The described technique for aspiration from the PSIS is safe, reliable, reproducible, and substantiated by data.

The iliac crest is an effective site for harvesting bone marrow stem cells. It allows for easy access and is superficial in most individuals, allowing for a relatively quick and simple procedure. Use of mesenchymal stem cells (MSCs) for treatment of orthopedic injuries has grown recently. Whereas overall use has increased, review of the literature reveals very few techniques for iliac crest aspiration,¹ but these are not based on anatomic relationships or studies. Hernigou and colleagues^2,3 attempted to quantitatively evaluate potential “sectors” allowing for safe aspiration using cadaver and computed tomographic reconstruction imaging. We used magnetic resonance imaging (MRI) to analyze aspiration parameters. Owing to the ilium’s anatomy, improper positioning or aspiration technique during aspiration can result in serious injury.^2,4-6 We hypothesized that there is an ideal angle and positioning for bone marrow aspiration from the posterior superior iliac spine (PSIS) that is safe, consistent, and reproducible. Although most aspiration techniques use landmark palpation, this is unreliable and inaccurate, especially when compared with ultrasound-guided injections^7-16 and procedures.^9,12,17-19 We describe our technique using ultrasound to visualize patient anatomy and accurately determine anatomic entry with the trocar.

METHODS

MRI scans of 26 patients (13 males, 13 females) were reviewed to determine average angles and distances. Axial T2-weighted views of the lumbar spine were used in all analyses. The sacroiliac (SI) joint angle was defined as the angle formed between the vector through the midline of the pelvis and the vector that is parallel to the SI joint. The approach angle was defined as the angle formed between the vector of the most medial aspect of the PSIS through the ilium to the anterior wall and the vector through the midline of the pelvis (Figure 1). 

The distance, in centimeters, of the PSIS to the anterior ilium wall was measured to determine the maximum distance the trocar can advance without puncturing the anterior ilium wall. The perpendicular distance from the PSIS table to the anterior aspect of the ilium was measured to determine the average depth the trocar could advance if the clinician entered it perpendicular to the patient’s coronal plane (Figure 2). 

The PSIS table width was defined as the width, in centimeters, of the PSIS posteriorly. The minimum width, in centimeters, of the ilium was measured at the thinnest region of the ilium to determine the degree of variation if the trocar were entered too far laterally and advanced to the anterior ilium wall (Figure 3). 

Means and standard deviations were calculated by sex and as a total. Student’s t test was used to determine statistical significance (P < .05).

Continue to: For the 13 males, the mean SI joint...

 

RESULTS

The results are reported in the Table. 

Table. Measurements of Patients Taken on Axial T2-Weighted Views of Lumbosacral MRI Scans^a

Patient	SI Joint Angle (°)	Approach Angle (°)	PSIS Table Width (cm)	PSIS to Anterior Ilium Wall (cm)	Perpendicular Distance PSIS to Anterior Joint (cm)	Post Ilium Wall to SI Joint Width (cm)
Males
1	28.80	19.50	1.24	8.80	4.16	1.52
2	31.80	27.60	1.70	7.89	3.49	1.02
3	33.70	27.70	1.12	8.14	3.15	1.28
4	23.70	26.40	0.95	6.66	3.22	0.65
5	35.90	28.40	0.84	7.60	2.57	0.95
6	33.80	29.30	1.20	7.73	2.34	0.90
7	30.30	21.20	1.36	8.44	3.95	1.18
8	34.50	20.40	1.53	7.08	3.98	1.56
9	28.70	24.00	1.34	8.19	3.51	1.31
10	22.40	20.10	1.37	7.30	3.87	1.28
11	33.60	20.80	0.88	6.43	3.26	0.94
12	48.50	31.00	1.15	6.69	2.97	1.38
13	20.20	20.90	0.94	6.95	3.79	1.05
Averages	31.22	24.41	1.20	7.53	3.40	1.16
Standard Deviation	7.18	4.11	0.26	0.75	0.56	0.26
Females
14	22.80	23.20	1.54	7.21	3.45	1.39
15	33.30	21.40	1.09	7.26	3.57	0.98
16	19.70	15.60	0.78	8.32	3.76	0.86
17	17.50	15.60	0.61	7.57	3.37	1.03
18	48.20	26.60	0.94	6.62	3.16	0.71
19	38.20	28.30	0.90	6.32	2.23	0.91
20	44.50	31.70	0.99	6.19	3.06	0.76
21	24.10	18.00	0.92	6.99	3.23	0.71
22	17.20	14.80	0.81	6.00	2.81	1.13
23	42.00	38.50	1.00	5.33	2.47	1.42
24	32.00	25.50	0.98	6.01	2.79	1.21
25	24.70	24.80	0.87	6.09	2.79	1.02
26	19.80	22.30	1.04	7.71	2.37	1.36
Averages	29.54	23.56	0.96	6.74	3.00	1.04
Standard Deviation	10.84	6.88	0.21	0.85	0.48	0.25
All patients Averages	30.38	23.98	1.08	7.14	3.20	1.10
Standard Deviation	9.05	5.57	0.26	0.88	0.55	0.26

^aStatistical significance is denoted as P < .02.

Abbreviations: MRI, magnetic resonance imaging; PSIS, posterior iliac spine; SI, sacroiliac.

For the 13 males, the mean SI joint angle was 31.22° ± 7.18° (range, 20.20° to 48.50°). The mean approach angle was 24.41° ± 4.11° (range, 19.50° to 31.00°). The mean PSIS table width was 1.20 cm ± 0.26 cm (range, 0.84 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.53 cm ± 0.75 cm (range, 6.43 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.40 cm ± 0.56 cm (range, 2.34 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.16 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

For the 13 females, the mean SI joint angle was 29.54° ± 10.84° (range, 17.20° to 48.20°). The mean approach angle was 23.56° ± 6.88° (range, 14.80° to 38.50°). The mean PSIS table width was 0.96 cm ± 0.21 cm (range, 0.61 cm to 1.54 cm). The mean distance from the PSIS to the anterior ilium wall was 6.74 cm ± 0.85 cm (range, 5.33 cm to 8.32 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.00 cm ± 0.48 cm (range, 2.23 cm to 3.76 cm). The mean minimum width of the ilium to the SI joint was 1.04 cm ± 0.25 cm (range, 0.71 cm to 1.42 cm).

For the 26 total patients, the mean SI joint angle was 30.38° ± 9.05° (range, 17.20° to 48.50°). The mean approach angle was 23.98° ± 5.57° (range, 14.80° to 38.50°). The mean PSIS table width was 1.08 cm ± 0.26 cm (range, 0.61 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.14 cm ± 0.88 cm (range, 5.33 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.20 cm ± 0.55 cm (range, 2.23 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.10 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

There was a statistically significant difference between the male and female groups for the maximum distance the trocar can be advanced from the PSIS to the anterior ilium wall (P < .02), and a statistically significant difference for the PSIS table width (P < .02). There were no significant differences between the male and female groups for the approach angle, the SI joint angle, the perpendicular distance from the PSIS to the anterior ilium, and the minimum width of the ilium to the SI joint.

Continue to: The patient is brought to the procedure...

 

TECHNIQUE: ILIAC CREST (PSIS) BONE MARROW ASPIRATION

The patient is brought to the procedure room and placed in a prone position. The donor site is prepared and draped in the usual sterile manner. Ultrasound is used to identify the median sacral crest in a short-axis view. The probe is then moved laterally to identify the PSIS (Figures 4A, 4B). 

The probe can be moved superiorly and inferiorly to determine the most prominent and central portion of the PSIS. The SI joint and ilium can also be visualized if needed.

The crosshairs on the ultrasound probe are used to mark the center lines of each plane. The central point marks the location of the PSIS. Alternatively, an in-plane technique can be used to place a spinal needle on the exact entry point on the PSIS. Once the PSIS and entry point are identified, the site is blocked with 10 mL of 0.5% ropivacaine.

Prior to introduction of the trocar, all instrumentation is primed with heparin and syringes are prepped with anticoagulant citrate dextrose solution, solution A. A stab incision is made at the site. The trocar is placed at the entry point, which should be centered in a superior-inferior plane and at the most medial point of the PSIS. Starting with the trocar vertical, the trocar is angled laterally 24° by dropping the hand medially toward the midline. No angulation cephalad or caudad is necessary, but cephalad must be avoided so as not to skive superiorly. This angle, which is recommended for both males and females, allows for the greatest distance the trocar can travel in bone before hitting the anterior ilium wall. A standard deviation of 5.57° is present, which should be considered. Steady pressure should be applied with a slight twisting motion on the PSIS. If advancement of the trocar is too difficult, a mallet or drill can be used to assist in penetration.

With the trocar advanced into the bone 1 cm, the trocar needle is removed while the cannula remains in place. The syringe is attached to the top of the cannula. The syringe plunger is pulled back to aspirate 20 mL of bone marrow. The cannula and syringe assembly are advanced 2 cm farther into the bone to allow for aspiration of a new location within the bone marrow cavity, and 20 mL of bone marrow are again aspirated. This is done a final time, advancing the trocar another 2 cm and aspirating a final 20 mL of bone marrow. The entire process should yield roughly 60 mL of bone marrow from one side. If desired, the same process can be repeated for the contralateral PSIS to yield a total of 120 mL of bone marrow from the 2 sites.

Based on our data, the average distance to the anterior ilium wall was 7 cm, but the shortest distance noted in this study was 5 cm. On the basis of the data presented, this technique allows for safe advancement based on even the shortest measured distance, without fear of puncturing the anterior ilium wall. Perforation could damage the femoral nerve and the internal or external iliac artery or vein that lie anterior to the ilium.

Continue to: We hypothesized that there...

 

DISCUSSION

We hypothesized that there would be an optimal angle of entry and maximal safe distance the trocar could advance through the ilium when aspirating. Because male and female pelvic anatomy differs, we also hypothesized that there would be differences in distance and size measurements for males and females. Our results supported our hypothesis that there is an ideal approach angle. The results also showed that the maximum distance the trocar can advance and the width of the PSIS table differ significantly between males and females.

Although pelvic anatomy differs between males and females, there should be an ideal entry angle that would allow maximum advancement into the ilium without perforating the anterior wall, which we defined as the approach angle. In our comparison of 26 MRI scans, we found that the approach angle did not differ significantly between the 2 groups (13 males, 13 females). This allows clinicians to enter the PSIS at roughly 24° medial to the parasagittal line, maximizing the space before puncturing into the anterior pelvis in either males or females.

If clinicians were to enter perpendicular to the patient’s PSIS, they would, on average, be able to advance only 3.20 cm before encountering the SI joint. When entering at 24° as we recommend, the average distance increases to 7.14 cm. Although the angle did not differ significantly, there was a significant difference between males and females in the length from the PSIS to the anterior wall, with males having 7.53 cm distance and females 6.74 cm. This is an important measurement because if the anterior ilium wall is punctured, the femoral nerve and the common, internal and external iliac arteries and veins could be damaged, resulting in retroperitoneal hemorrhage.

A fatality in 2001 in the United Kingdom led to a national audit of bone marrow aspiration and biopsies.^4-6 Although these procedures were done primarily for patients with cancer, hemorrhagic events were the most frequent and serious events. This audit led to the identification of many risk factors. Bain^4-6 conducted reviews of bone marrow aspirations and biopsies in the United Kingdom from 2002 to 2004. Of a total of 53,088 procedures conducted during that time frame, 48 (0.09%) adverse events occurred, with 29 (0.05%) being hemorrhagic events. Although infrequent, hemorrhagic adverse events represent significant morbidity. Reviews such as those conducted by Bain^4-6 highlight the importance of a study that helps determine the optimal parameters for aspiration to ensure safety and reliability.

Hernigou and colleagues^2,3 conducted studies analyzing different “sectors” in an attempt to develop a safe aspiration technique. They found that obese patients were at higher risk, and some sites of aspiration (sectors 1, 4, 5) had increased risk for perforation and damage to surrounding structures. Their sector 6, which incorporated the entirety of the PSIS table, was considered the safest, most reliable site for trocar introduction.^2,3 Hernigou and colleagues,² in comparing the bone mass of the sectors, also noted that sector 6 has the greatest bone thickness close to the entry point, making it the most favorable site. The PSIS is not just a point; it is more a “table.” The PSIS can be palpated posteriorly, but this is inaccurate and unreliable, particularly in larger individuals. The PSIS table can be identified on ultrasound before introducing the trocar, which is a more reliable method of landmark identification than palpation guidance, just as in ultrasound-guided injections^7-16 and procedures.^9,12,17-19

Continue to: If the PSIS is not accurately...

 

If the PSIS is not accurately identified, penetration laterally will result in entering the ilium wing, where it is quite narrow. We found the distance between the posterior ilium wall and the SI joint to be only 1.10 cm wide (Figure 3); we defined this area as the narrow corridor. Superior and lateral entry could damage the superior cluneal nerves coming over the iliac crest, which are located 6 cm lateral to the SI joint. Inferior and lateral entry 6 cm below the PSIS could reach the greater sciatic foramen, damaging the sacral plexus and superior gluteal artery and vein. If the entry slips above the PSIS over the pelvis, the trocar could enter the retroperitoneal space and damage the femoral nerve and common iliac artery and vein, leading to a retroperitoneal hemorrhage.^4-6,20

MSCs are found as perivascular cells and lie in the cortices of bones.²¹ Following the approach angle and directed line from the PSIS to the anterior ilium wall described in this study (Figures 1 and 2), the trocar would pass through the narrow corridor as it advances farther into the ilium. The minimum width of this corridor was measured in this study and, on average, was 1.10 cm wide from cortex to cortex (Figure 3). As the bone marrow is aspirated from this narrow corridor, the clinician is gathering MSCs from both the lateral and medial cortices of the ilium. By aspirating from a greater surface area of the cortices, it is believed that this will increase the total collection of MSCs.

CONCLUSION

Although there are reports in the literature that describe techniques for bone marrow aspiration from the iliac crest, the techniques are very general and vague regarding the ideal angles and methods. Studies have attempted to quantify the safest entry sites for aspiration but have not detailed ideal parameters for collection. Blind aspiration from the iliac crest can have serious implications if adverse events occur, and thus there is a need for a safe and reliable method of aspiration from the iliac crest. Ultrasound guidance to identify anatomy, as opposed to palpation guidance, ensures anatomic placement of the trocar while minimizing the risk of aspiration. Based on the measurements gathered in this study, an optimal angle of entry and safe distance of penetration have been identified. Using our data and relevant literature, we developed a technique for a safe, consistent, and reliable method of bone marrow aspiration out of the iliac crest.

ABSTRACT

Use of mesenchymal stem cells from bone marrow has gained significant popularity. The iliac crest has been determined to be an effective site for harvesting mesenchymal stem cells. Review of the literature reveals that multiple techniques are used to harvest bone marrow aspirate from the iliac crest, but the descriptions are based on the experience of various authors as opposed to studied anatomy. A safe, reliable, and reproducible method for aspiration has yet to be studied and described. We hypothesized that there would be an ideal angle and distance for aspiration that would be the safest, most consistent, and most reliable. Using magnetic resonance imaging (MRI), we reviewed 26 total lumbar spine MRI scans (13 males, 13 females) and found that an angle of 24° should be used when entering the most medial aspect of the posterior superior iliac spine (PSIS) and that this angle did not differ between the sexes. The distance that the trocar can advance after entry before hitting the anterior ilium wall varied significantly between males and females, being 7.53 cm in males and 6.74 cm in females. In addition, the size of the PSIS table was significantly different between males and females (1.20 cm and 0.96 cm, respectively). No other significant differences in the measurements gathered were found. Using the data gleaned from this study, we developed an aspiration technique. This method uses ultrasound to determine the location of the PSIS and the entry point on the PSIS. This contrasts with most techniques that use landmark palpation, which is known to be unreliable and inaccurate. The described technique for aspiration from the PSIS is safe, reliable, reproducible, and substantiated by data.

The iliac crest is an effective site for harvesting bone marrow stem cells. It allows for easy access and is superficial in most individuals, allowing for a relatively quick and simple procedure. Use of mesenchymal stem cells (MSCs) for treatment of orthopedic injuries has grown recently. Whereas overall use has increased, review of the literature reveals very few techniques for iliac crest aspiration,¹ but these are not based on anatomic relationships or studies. Hernigou and colleagues^2,3 attempted to quantitatively evaluate potential “sectors” allowing for safe aspiration using cadaver and computed tomographic reconstruction imaging. We used magnetic resonance imaging (MRI) to analyze aspiration parameters. Owing to the ilium’s anatomy, improper positioning or aspiration technique during aspiration can result in serious injury.^2,4-6 We hypothesized that there is an ideal angle and positioning for bone marrow aspiration from the posterior superior iliac spine (PSIS) that is safe, consistent, and reproducible. Although most aspiration techniques use landmark palpation, this is unreliable and inaccurate, especially when compared with ultrasound-guided injections^7-16 and procedures.^9,12,17-19 We describe our technique using ultrasound to visualize patient anatomy and accurately determine anatomic entry with the trocar.

METHODS

MRI scans of 26 patients (13 males, 13 females) were reviewed to determine average angles and distances. Axial T2-weighted views of the lumbar spine were used in all analyses. The sacroiliac (SI) joint angle was defined as the angle formed between the vector through the midline of the pelvis and the vector that is parallel to the SI joint. The approach angle was defined as the angle formed between the vector of the most medial aspect of the PSIS through the ilium to the anterior wall and the vector through the midline of the pelvis (Figure 1). 

The distance, in centimeters, of the PSIS to the anterior ilium wall was measured to determine the maximum distance the trocar can advance without puncturing the anterior ilium wall. The perpendicular distance from the PSIS table to the anterior aspect of the ilium was measured to determine the average depth the trocar could advance if the clinician entered it perpendicular to the patient’s coronal plane (Figure 2). 

The PSIS table width was defined as the width, in centimeters, of the PSIS posteriorly. The minimum width, in centimeters, of the ilium was measured at the thinnest region of the ilium to determine the degree of variation if the trocar were entered too far laterally and advanced to the anterior ilium wall (Figure 3). 

Means and standard deviations were calculated by sex and as a total. Student’s t test was used to determine statistical significance (P < .05).

Continue to: For the 13 males, the mean SI joint...

 

RESULTS

The results are reported in the Table. 

Table. Measurements of Patients Taken on Axial T2-Weighted Views of Lumbosacral MRI Scans^a

Patient	SI Joint Angle (°)	Approach Angle (°)	PSIS Table Width (cm)	PSIS to Anterior Ilium Wall (cm)	Perpendicular Distance PSIS to Anterior Joint (cm)	Post Ilium Wall to SI Joint Width (cm)
Males
1	28.80	19.50	1.24	8.80	4.16	1.52
2	31.80	27.60	1.70	7.89	3.49	1.02
3	33.70	27.70	1.12	8.14	3.15	1.28
4	23.70	26.40	0.95	6.66	3.22	0.65
5	35.90	28.40	0.84	7.60	2.57	0.95
6	33.80	29.30	1.20	7.73	2.34	0.90
7	30.30	21.20	1.36	8.44	3.95	1.18
8	34.50	20.40	1.53	7.08	3.98	1.56
9	28.70	24.00	1.34	8.19	3.51	1.31
10	22.40	20.10	1.37	7.30	3.87	1.28
11	33.60	20.80	0.88	6.43	3.26	0.94
12	48.50	31.00	1.15	6.69	2.97	1.38
13	20.20	20.90	0.94	6.95	3.79	1.05
Averages	31.22	24.41	1.20	7.53	3.40	1.16
Standard Deviation	7.18	4.11	0.26	0.75	0.56	0.26
Females
14	22.80	23.20	1.54	7.21	3.45	1.39
15	33.30	21.40	1.09	7.26	3.57	0.98
16	19.70	15.60	0.78	8.32	3.76	0.86
17	17.50	15.60	0.61	7.57	3.37	1.03
18	48.20	26.60	0.94	6.62	3.16	0.71
19	38.20	28.30	0.90	6.32	2.23	0.91
20	44.50	31.70	0.99	6.19	3.06	0.76
21	24.10	18.00	0.92	6.99	3.23	0.71
22	17.20	14.80	0.81	6.00	2.81	1.13
23	42.00	38.50	1.00	5.33	2.47	1.42
24	32.00	25.50	0.98	6.01	2.79	1.21
25	24.70	24.80	0.87	6.09	2.79	1.02
26	19.80	22.30	1.04	7.71	2.37	1.36
Averages	29.54	23.56	0.96	6.74	3.00	1.04
Standard Deviation	10.84	6.88	0.21	0.85	0.48	0.25
All patients Averages	30.38	23.98	1.08	7.14	3.20	1.10
Standard Deviation	9.05	5.57	0.26	0.88	0.55	0.26

^aStatistical significance is denoted as P < .02.

Abbreviations: MRI, magnetic resonance imaging; PSIS, posterior iliac spine; SI, sacroiliac.

For the 13 males, the mean SI joint angle was 31.22° ± 7.18° (range, 20.20° to 48.50°). The mean approach angle was 24.41° ± 4.11° (range, 19.50° to 31.00°). The mean PSIS table width was 1.20 cm ± 0.26 cm (range, 0.84 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.53 cm ± 0.75 cm (range, 6.43 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.40 cm ± 0.56 cm (range, 2.34 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.16 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

For the 13 females, the mean SI joint angle was 29.54° ± 10.84° (range, 17.20° to 48.20°). The mean approach angle was 23.56° ± 6.88° (range, 14.80° to 38.50°). The mean PSIS table width was 0.96 cm ± 0.21 cm (range, 0.61 cm to 1.54 cm). The mean distance from the PSIS to the anterior ilium wall was 6.74 cm ± 0.85 cm (range, 5.33 cm to 8.32 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.00 cm ± 0.48 cm (range, 2.23 cm to 3.76 cm). The mean minimum width of the ilium to the SI joint was 1.04 cm ± 0.25 cm (range, 0.71 cm to 1.42 cm).

For the 26 total patients, the mean SI joint angle was 30.38° ± 9.05° (range, 17.20° to 48.50°). The mean approach angle was 23.98° ± 5.57° (range, 14.80° to 38.50°). The mean PSIS table width was 1.08 cm ± 0.26 cm (range, 0.61 cm to 1.70 cm). The mean distance from the PSIS to the anterior ilium wall was 7.14 cm ± 0.88 cm (range, 5.33 cm to 8.80 cm). The mean perpendicular distance from the PSIS table to the anterior ilium was 3.20 cm ± 0.55 cm (range, 2.23 cm to 4.16 cm). The mean minimum width of the ilium to the SI joint was 1.10 cm ± 0.26 cm (range, 0.65 cm to 1.56 cm).

There was a statistically significant difference between the male and female groups for the maximum distance the trocar can be advanced from the PSIS to the anterior ilium wall (P < .02), and a statistically significant difference for the PSIS table width (P < .02). There were no significant differences between the male and female groups for the approach angle, the SI joint angle, the perpendicular distance from the PSIS to the anterior ilium, and the minimum width of the ilium to the SI joint.

Continue to: The patient is brought to the procedure...

 

TECHNIQUE: ILIAC CREST (PSIS) BONE MARROW ASPIRATION

The patient is brought to the procedure room and placed in a prone position. The donor site is prepared and draped in the usual sterile manner. Ultrasound is used to identify the median sacral crest in a short-axis view. The probe is then moved laterally to identify the PSIS (Figures 4A, 4B). 

The probe can be moved superiorly and inferiorly to determine the most prominent and central portion of the PSIS. The SI joint and ilium can also be visualized if needed.

The crosshairs on the ultrasound probe are used to mark the center lines of each plane. The central point marks the location of the PSIS. Alternatively, an in-plane technique can be used to place a spinal needle on the exact entry point on the PSIS. Once the PSIS and entry point are identified, the site is blocked with 10 mL of 0.5% ropivacaine.

Prior to introduction of the trocar, all instrumentation is primed with heparin and syringes are prepped with anticoagulant citrate dextrose solution, solution A. A stab incision is made at the site. The trocar is placed at the entry point, which should be centered in a superior-inferior plane and at the most medial point of the PSIS. Starting with the trocar vertical, the trocar is angled laterally 24° by dropping the hand medially toward the midline. No angulation cephalad or caudad is necessary, but cephalad must be avoided so as not to skive superiorly. This angle, which is recommended for both males and females, allows for the greatest distance the trocar can travel in bone before hitting the anterior ilium wall. A standard deviation of 5.57° is present, which should be considered. Steady pressure should be applied with a slight twisting motion on the PSIS. If advancement of the trocar is too difficult, a mallet or drill can be used to assist in penetration.

With the trocar advanced into the bone 1 cm, the trocar needle is removed while the cannula remains in place. The syringe is attached to the top of the cannula. The syringe plunger is pulled back to aspirate 20 mL of bone marrow. The cannula and syringe assembly are advanced 2 cm farther into the bone to allow for aspiration of a new location within the bone marrow cavity, and 20 mL of bone marrow are again aspirated. This is done a final time, advancing the trocar another 2 cm and aspirating a final 20 mL of bone marrow. The entire process should yield roughly 60 mL of bone marrow from one side. If desired, the same process can be repeated for the contralateral PSIS to yield a total of 120 mL of bone marrow from the 2 sites.

Based on our data, the average distance to the anterior ilium wall was 7 cm, but the shortest distance noted in this study was 5 cm. On the basis of the data presented, this technique allows for safe advancement based on even the shortest measured distance, without fear of puncturing the anterior ilium wall. Perforation could damage the femoral nerve and the internal or external iliac artery or vein that lie anterior to the ilium.

Continue to: We hypothesized that there...

 

DISCUSSION

We hypothesized that there would be an optimal angle of entry and maximal safe distance the trocar could advance through the ilium when aspirating. Because male and female pelvic anatomy differs, we also hypothesized that there would be differences in distance and size measurements for males and females. Our results supported our hypothesis that there is an ideal approach angle. The results also showed that the maximum distance the trocar can advance and the width of the PSIS table differ significantly between males and females.

Although pelvic anatomy differs between males and females, there should be an ideal entry angle that would allow maximum advancement into the ilium without perforating the anterior wall, which we defined as the approach angle. In our comparison of 26 MRI scans, we found that the approach angle did not differ significantly between the 2 groups (13 males, 13 females). This allows clinicians to enter the PSIS at roughly 24° medial to the parasagittal line, maximizing the space before puncturing into the anterior pelvis in either males or females.

If clinicians were to enter perpendicular to the patient’s PSIS, they would, on average, be able to advance only 3.20 cm before encountering the SI joint. When entering at 24° as we recommend, the average distance increases to 7.14 cm. Although the angle did not differ significantly, there was a significant difference between males and females in the length from the PSIS to the anterior wall, with males having 7.53 cm distance and females 6.74 cm. This is an important measurement because if the anterior ilium wall is punctured, the femoral nerve and the common, internal and external iliac arteries and veins could be damaged, resulting in retroperitoneal hemorrhage.

A fatality in 2001 in the United Kingdom led to a national audit of bone marrow aspiration and biopsies.^4-6 Although these procedures were done primarily for patients with cancer, hemorrhagic events were the most frequent and serious events. This audit led to the identification of many risk factors. Bain^4-6 conducted reviews of bone marrow aspirations and biopsies in the United Kingdom from 2002 to 2004. Of a total of 53,088 procedures conducted during that time frame, 48 (0.09%) adverse events occurred, with 29 (0.05%) being hemorrhagic events. Although infrequent, hemorrhagic adverse events represent significant morbidity. Reviews such as those conducted by Bain^4-6 highlight the importance of a study that helps determine the optimal parameters for aspiration to ensure safety and reliability.

Hernigou and colleagues^2,3 conducted studies analyzing different “sectors” in an attempt to develop a safe aspiration technique. They found that obese patients were at higher risk, and some sites of aspiration (sectors 1, 4, 5) had increased risk for perforation and damage to surrounding structures. Their sector 6, which incorporated the entirety of the PSIS table, was considered the safest, most reliable site for trocar introduction.^2,3 Hernigou and colleagues,² in comparing the bone mass of the sectors, also noted that sector 6 has the greatest bone thickness close to the entry point, making it the most favorable site. The PSIS is not just a point; it is more a “table.” The PSIS can be palpated posteriorly, but this is inaccurate and unreliable, particularly in larger individuals. The PSIS table can be identified on ultrasound before introducing the trocar, which is a more reliable method of landmark identification than palpation guidance, just as in ultrasound-guided injections^7-16 and procedures.^9,12,17-19

Continue to: If the PSIS is not accurately...

 

If the PSIS is not accurately identified, penetration laterally will result in entering the ilium wing, where it is quite narrow. We found the distance between the posterior ilium wall and the SI joint to be only 1.10 cm wide (Figure 3); we defined this area as the narrow corridor. Superior and lateral entry could damage the superior cluneal nerves coming over the iliac crest, which are located 6 cm lateral to the SI joint. Inferior and lateral entry 6 cm below the PSIS could reach the greater sciatic foramen, damaging the sacral plexus and superior gluteal artery and vein. If the entry slips above the PSIS over the pelvis, the trocar could enter the retroperitoneal space and damage the femoral nerve and common iliac artery and vein, leading to a retroperitoneal hemorrhage.^4-6,20

MSCs are found as perivascular cells and lie in the cortices of bones.²¹ Following the approach angle and directed line from the PSIS to the anterior ilium wall described in this study (Figures 1 and 2), the trocar would pass through the narrow corridor as it advances farther into the ilium. The minimum width of this corridor was measured in this study and, on average, was 1.10 cm wide from cortex to cortex (Figure 3). As the bone marrow is aspirated from this narrow corridor, the clinician is gathering MSCs from both the lateral and medial cortices of the ilium. By aspirating from a greater surface area of the cortices, it is believed that this will increase the total collection of MSCs.

CONCLUSION

Although there are reports in the literature that describe techniques for bone marrow aspiration from the iliac crest, the techniques are very general and vague regarding the ideal angles and methods. Studies have attempted to quantify the safest entry sites for aspiration but have not detailed ideal parameters for collection. Blind aspiration from the iliac crest can have serious implications if adverse events occur, and thus there is a need for a safe and reliable method of aspiration from the iliac crest. Ultrasound guidance to identify anatomy, as opposed to palpation guidance, ensures anatomic placement of the trocar while minimizing the risk of aspiration. Based on the measurements gathered in this study, an optimal angle of entry and safe distance of penetration have been identified. Using our data and relevant literature, we developed a technique for a safe, consistent, and reliable method of bone marrow aspiration out of the iliac crest.

The majority of hospitalized patients worldwide have at least one peripheral intravenous catheter (PIVC),¹ making PIVC insertion one of the most common clinical procedures. In the United States, physicians, advanced practitioners, and nurses insert over 300 million of these devices in hospitalized patients annually.² Despite their prevalence, PIVCs are associated with high rates of complications, including insertion difficulty, phlebitis, infiltration, occlusion, dislodgment, and catheter-associated bloodstream infection (CABSI), known to increase morbidity and mortality risk.^2-9 Up to 90% of PIVCs are prematurely removed owing to failure before planned replacement or before intravenous (IV) therapy completion.^3-6,10-12

PIVC complication and failure commonly triggers insertion of a replacement device and can entail significant costs.^2-4 One example is PIVC-related CABSI, where treatment costs have been estimated to be between US$35,000 and US$56,000 per patient.^6,13 Another important consideration is the pain and anxiety experienced by patients who need a replacement device, particularly those with difficult vascular access, who may require multiple cannulation attempts to replace a PIVC.^12,14-16 In developing nations, serious adverse events related to PIVCs are even more concerning, because hospital acquired infection rates and associated mortality are nearly 20 times greater than in developed nations.¹⁷

A number of evidence-based interventions have been suggested to reduce PIVC failure rates. In addition to optimal hand hygiene when inserting or accessing a PIVC to prevent infection,¹⁸ recommended interventions include placement of the PIVC in an area of non-flexion such as the forearm to provide stability for the device and to reduce patient discomfort, securing the PIVC to reduce movement of the catheter at the insertion site and within the blood vessel, and use of occlusive dressings that reduce the risk of external contamination of the PIVC site.^11,19,20 Best practice guidelines also recommend the prompt removal of devices that are symptomatic (when phlebitis or other complications are suspected) and when the catheter is no longer required.^21,22

Recent evidence has demonstrated that catheter size can have an impact on device survival rates. In adults, large-bore catheters of 18 gauge (G) or higher were found to have an increased rate of thrombosis, and smaller-bore catheters of 22G or lower (in adults) were found to have higher rates of dislodgment and occlusion/infiltration. The catheter size recommended for adults based on the latest evidence for most clinical applications is 20G.^3,20,23,24 In addition, the documentation of insertion, maintenance, and removal of PIVCs in the medical record is a requirement in most healthcare facilities worldwide and is recommended by best practice guidelines; however, adherence remains a challenge.^1,19

The concerning prevalence of PIVC-related complications and the lack of comparative data internationally on organizational compliance with best practice guidelines formed the rationale for this study. Our study aim was to describe the insertion characteristics, management practices, and outcomes of PIVCs internationally and to compare these variables to recommended best practice.

MATERIALS AND METHODS

Study Design and Participants

In this international cross-sectional study, we recruited hospitals through professional networks, including vascular access, infection prevention, safety and quality, nursing, and hospital associations (Appendix 2). Healthcare organizations, government health departments, and intravascular device suppliers were informed of the study and requested to further disseminate information through their networks. A study website was developed,²⁵ and social media outlets, including Twitter^®, LinkedIn^®, and Facebook^®, were used to promote the study.

 

Approval was granted by the Griffith University Human Research Ethics Committee in Australia (reference number NRS/34/13/HREC). In addition, evidence of study site and local institutional review board/ethics committee approval was required prior to study commencement. Each participating site agreed to follow the study protocol and signed an authorship agreement form. No financial support was provided to any site.

Hospitalized adult and pediatric patients with a PIVC in situ on the day of the study were eligible for inclusion. Sample size was determined by local capacity. Hospitals were encouraged to audit their entire institution if possible; however, data were accepted from as little as one ward. Data collectors comprised nurses and doctors with experience in PIVC assessment. They were briefed on the study protocol and data collection forms by the local site coordinator, and they were supported by an overall global coordinator. Clinicians assessed the PIVC insertion site and accessed hospital records to collect data related to PIVC insertion, concurrent medications, and IV fluid orders. Further clarification of data was obtained if necessary by the clinicians from the patients and treating staff. No identifiable patient information was collected.

Data Collection

To assess whether clinical facilities were following best practice recommendations, the study team developed three data collection forms to collect information regarding site characteristics (site questionnaire), track participant recruitment (screening log), and collect data regarding PIVC characteristics and management practices (case report form [CRF]). All forms were internally and externally validated following a pilot study involving 14 sites in 13 countries.¹

The CRF included variables used to assess best practice interventions, such as catheter insertion characteristics (date and time, reason, location, profession of inserter, anatomical site of placement), catheter type (gauge, brand, and product), insertion site assessment (adverse symptoms, dressing type and integrity), and information related to the IV therapy (types of IV fluids and medications, flushing solutions). Idle PIVCs were defined as not being used for blood sampling or IV therapy in the preceding 24 h.

Data collection forms were translated into 15 languages by professional translators and back-translated for validity. Translation of some languages included additional rigor. For example, Spanish-speaking members from the Spanish mainland as well as from South America were employed so that appropriate synonyms were used to capture local terms and practice. Three options were provided for data entry: directly into a purpose-developed electronic database (Lime Survey^® Project, Hamburg, Germany); on paper, then transcribed into the survey database at a later time by the hospital site; or paper entry then sent (via email or post) to the coordinating center for data entry. Once cleaned and collated, all data were provided to each participating hospital to confirm accuracy and for site use in local quality improvement processes. Data were collected between June 1, 2014 and July 31, 2015.

Statistical Analysis

All data management was undertaken using SAS statistical software (SAS Institute Inc., Cary NC, USA). Results are presented for eight geographical regions using descriptive statistics (frequencies, percentages, and 95% CIs) for the variables of interest. To assess trends in catheter dwell time and rates of phlebitis, Poisson regression was used. All analyses were undertaken using the R language for statistical analysis (R Core Team, Vienna, Austria). The (STROBE (Strengthening the Reporting of Observational Studies in Epidemiology statement) guidelines for cross-sectional studies were followed, and results are presented according to these recommendations.²⁶

RESULTS

Of the 415 hospitals that participated in this study, 406 had patients with PIVCs on the day of the study (the others being small rural centers). Thus, a total of 40,620 PIVCs in 38,161 patients from 406 hospitals in 51 countries were assessed, with no more than 5% missing data for any CRF question. There were 2459 patients (6.1%) with two or more PIVCs concurrently in situ. The median patient age was 59 y (interquartile range [IQR], 37–74 y), and just over half were male (n = 20,550, 51%). Hospital size ranged from fewer than 10 beds to over 1,000 beds, and hospitals were located in rural, regional, and metropolitan districts. The majority of countries (n = 31, 61%) contributed multiple sites, the highest being Australia with 79 hospitals. Countries with the most PIVCs studied were Spain (n = 5,553, 14%) and the United States (n = 5,048, 12%).

General surgical (n = 15,616, 39%) and medical (n = 15,448, 38%) patients represented most of the population observed. PIVCs were inserted primarily in general wards or clinics (n = 22,167, 55%) or in emergency departments (n = 7,388, 18%; Table) and for the administration of IV medication (n = 28,571, 70%) and IV fluids (n = 7,093, 18%; Table).

Globally, nurses were the primary PIVC inserters (n = 28,575, 71%); however, Australia/New Zealand had only 26% (n = 1,518) of PIVCs inserted by this group (Table). Only about one-third of PIVCs were placed in an area of non-flexion (forearm, n = 12,675, 31%, Table) the majority (n = 27,856, 69%) were placed in non-recommended anatomical sites (Figure 1). Most PIVCs were placed in the hand (n = 13,265, 32.7%) followed by the antecubital veins (n = 6176, 15.2%) and the wrist (n = 5,465, 13.5%). Site selection varied widely across the regions; 29% (n = 1686) of PIVCs in Australia/New Zealand were inserted into the antecubital veins, twice the study group average. Over half of the PIVCs inserted in the Middle East were placed in the hand (n = 295, 56%). This region also had the highest prevalence of devices placed in nonrecommended sites (n = 416, 79%; Figure 1).

The majority of PIVCs (n = 27,192, 67%; Table) were of recommended size (20–22G); however, some devices were observed to be large (14–18G; n = 6,802, 17%) or small (24-26g; n = 4,869, 12%) in adults. In Asia, 41% (n = 2,617) of devices inserted were 24-26G, more than three times the global rate. Half of all devices in Asia (n = 3,077, 48%) and the South Pacific (n = 67, 52%) were of a size not recommended for routine IV therapy (Figure 2).

The primary dressing material used was a transparent dressing (n = 31,596, 77.8%; Table); however, nearly 1 in 5 dressings used had either nonsterile tape alone (n = 5,169, 13%; Appendix 4), or a sterile gauze and tape (n = 2,592, 6%; Appendix 4.1). We found a wide variation in the use of nonsterile tape, including 1 in every 3 devices in South America dressed with nonsterile tape (n = 714, 30%) and a larger proportion in Africa (n = 543, 19%) and Europe (n = 3,056, 18%). Nonsterile tape was rarely used in North America and Australia/New Zealand. Although most PIVC dressings were clean, dry, and intact (n = 31,786, 79%; Table), one-fifth overall were compromised (moist, soiled, and/or lifting off the skin). Compromised dressings (Appendix 4.2) were more prevalent in Australia/New Zealand (n = 1,448; 25%) and in Africa (n = 707, 25%) than elsewhere.

Ten percent of PIVCs (n = 4,204) had signs and/or symptoms suggestive of phlebitis (characterized by pain, redness and/or swelling at the insertion site; Appendix 4.3). The highest prevalence of phlebitis occurred in Asia (n = 1,021, 16%), Africa (n = 360, 13%), and South America (n = 284, 12%). Pain and/or redness were the most common phlebitis symptoms. We found no association between dwell time of PIVCs and phlebitis rates (P = .085). Phlebitis rates were 12% (Days 1-3; n = 15,625), 16% (Days 4-7; n = 3,348), 10% (Days 8-21; n = 457), and 13% (Day21+; n = 174). Nearly 10% (n = 3,879) of catheters were observed to have signs of malfunction such as blood in the infusion tubing, leaking at the insertion site, or dislodgment (Appendix 4.4).

We observed 14% (n = 5,796) of PIVCs to be idle (Appendix 4.5), defined as not used in the preceding 24 h. Nearly one-fourth of all devices in North America (n = 1,230, 23%) and Australia/New Zealand (n = 1,335, 23%) were idle. PIVC documentation in hospital records was also poor, nearly half of all PIVCs (n = 19,768, 49%) had no documented date and time of insertion. The poorest compliance was in Australia/New Zealand (n = 3,428, 59%; Appendix 4.6). We also observed that 1 in 10 PIVCs had no documentation regarding who inserted the PIVC (n = 3,905). Thirty-six percent of PIVCs (n = 14,787) had no documented assessment of the PIVC site on the day of review (Appendix 4.7), including over half of all PIVCs in Asia (n = 3,364, 52%). Overall, the median dwell at the time of assessment for PIVCs with insertion date/time documented was 1.5 d (IQR, 1.0–2.5 d).

 

DISCUSSION

This international assessment of more than 40,000 PIVCs in 51 countries provides great insight into device characteristics and variation in management practices. Predominantly, PIVCs were inserted by nurses in the general ward environment for IV medication. One in ten PIVCs had at least one symptom of phlebitis, one in ten were dysfunctional, one in five PIVC dressings were compromised, and one in six PIVCs had not been used in the preceding 24 h. Nearly half of the PIVCs audited had the insertion date and time missing.

Regional variation was found in the professions inserting PIVCs, as well as in anatomical placement. In Australia/New Zealand, the proportion of nurses inserting PIVCs was much lower than the study group average (26% vs 71%). Because these countries contributed a substantial number of hospitals to the study, this seems a representative finding and suggests a need for education targeted at nurses for PIVC insertion in this region. The veins in the forearm are recommended as optimal for PIVC insertion in adults, rather than areas of high flexion, because the forearm provides a wide surface area to secure and dress PIVCs. Forearm placement can reduce pain during catheter dwell as well as decrease the risk of accidental removal or occlusion.^3,19,27 We found only one-third of PIVCs were placed in the forearm, with most placed in the hand, antecubital veins, or wrist. This highlights an inconsistency with published recommendations and suggests that additional training and technology are required so that staff can better identify and insert PIVCs in the forearm for other than very short-term (procedural) PIVCp;s.¹⁹

Phlebitis triggering PIVC failure remains a global clinical challenge with numerous phlebitis definitions and varied assessment techniques.¹⁰ The prevalence of phlebitis has been difficult to approximate with varying estimates and definitions in the literature; however, it remains a key predictor of PIVC failure.^6,10 Identification of this complication and prompt removal of the device is critical for patient comfort and reducing CABSI risk.^5,28 The overall prevalence of phlebitis signs or symptoms (defined in this study as having one or more signs of redness, swelling, or pain surrounding the insertion site) was just over 10%, with pain and/or redness being most prevalent. These compromised PIVCs had not been removed as is recommended for such complications.^19,28 Considering that our study was a snapshot at only one time point, the per-catheter incidence of phlebitis would be even higher; interestingly, among PIVCs with a documented insertion date and time, we observed that dwell time did not influence phlebitis rates.

Another concern is that nearly 10% (n = 3,879) of PIVCs were malfunctioning (eg, leaking) but were still in place. To bring these problems into context, around 2 billion PIVCs are used annually worldwide; as a consequence, millions of patients suffer from painful or malfunctioning PIVCs staff had not responded.^1,29 The placement of large-bore catheters, and smaller-gauge ones in adults, is known to increase the incidence of malfunction that leads to failure. There are a number of sound clinical reasons for the use of large-bore (eg, resuscitation and rapid fluid replacement) or small-bore (eg, difficult venous access with small superficial veins only visible and palpable) catheters. However, it would be expected that only a small proportion of patients would require these devices, and not one in three devices as we identified. This finding suggests that some PIVCs were inappropriate in size for general IV therapy and may reflect antiquated hospital policies for some clinical cohorts.^30,31

Overall, transparent dressings were used to cover the PIVC, but a number of patients were observed to have a sterile gauze and tape dressing (n = 2,592, 6%). Although the latter is less common, both dressing approaches are recommended in clinical practice guidelines because there is a lack of high-quality evidence regarding which is superior.^21,22,32 Of concern was the use of nonsterile tape to dress the PIVC (n = 5,169, 12.7%). We found the prevalence of nonsterile tape use to be higher in lower-resourced countries in South America (n = 714, 30%), Africa (n = 543, 19%) and Europe (n = 3,056, 18%) and this was likely related to institutional cost reduction practices.

This finding illustrates an important issue regarding proper PIVC care and management practices in developing nations. It is widely known that access to safe health care in lower-resourced nations is challenging and that rates of mortality related to healthcare-associated infections are much higher. Thus, the differences we found in PIVC management practices in these countries are not surprising.^33,34 International health networks such as the Infection Control Africa Network, the International Federation of Infection Control, and the Centers for Disease Control and Prevention can have great influence on ministries of health and clinicians in these countries to develop coordinated efforts for safe and sustainable IV practices to reduce the burden of hospital-acquired infections and related morbidity and mortality.

We found that 14% of all PIVCs had no documented IV medication or IV fluid administered in the previous 24 h, strongly indicating that they were no longer needed. Australia/New Zealand, Europe, and North America were observed to have a higher prevalence of idle catheters than the remaining regions. This suggests that an opportunity exists to develop surveillance systems that better identify idle devices for prompt removal to reduce infection risk and patient discomfort. Several randomized controlled trials, a Cochrane review, and clinical practice guidelines recommend prompt removal of PIVCs when not required, if there are any complications, or if the PIVC was inserted urgently without an aseptic insertion technique.^21,28,35,36 Idle PIVCs have been implicated in adverse patient outcomes, including phlebitis and CABSI.^13,27

The substantial proportion of patients with a PIVC in this study who had no clinical indication for a PIVC, a symptomatic insertion site, malfunctioning catheter, and suboptimal dressing quality suggests the need for physicians, advanced practitioners, and nurses to adopt evidence-based PIVC insertion and maintenance bundles and supporting checklists to reduce the prevalence of PIVC complications.^19,21,38-40 Recommended strategies for inclusion in PIVC maintenance bundles are prompt removal of symptomatic and/or idle catheters, hand hygiene prior to accessing the catheter, regular assessment of the device, and replacement of suboptimal dressings.^41,42 This approach should be implemented across all clinical specialties involved in PIVC insertion and care.

Our study findings need to be considered within the context of some limitations. The cross-sectional design prevented follow-up of PIVCs until removal to collect outcomes, including subsequent PIVC complications and/or failure, following the study observation. Ideally, data collection could have included patient-level preferences for PIVC insertion, history of PIVC use and/or failure, the number of PIVC insertion attempts, and the number of PIVCs used during that hospitalization. However, a cohort study of this magnitude was not feasible, particularly because all sites contributed staff time to complete the data collection. Only half of all initially registered sites eventually participated in the study; reasons for not participating were cited as local workload constraints and/or difficulties in applying for local approvals. Although efforts to enroll hospitals worldwide were exhaustive, our sample was not randomly selected but relied on self-selection and so is not representative, particularly for countries that contributed only one hospital site. Caution is also required when comparing inter regional differences, particularly developing regions, because better-resourced/academic sites were possibly over represented in the sample. Nevertheless, PIVC variables differed significantly between participating hospitals, suggesting that the data represent a reasonable reflection of hospital variability.

 

CONCLUSIONS

On the basis of this international investigation, we report variations in the characteristics, management practices, and outcomes of PIVCs inserted in hospital patients from 51 countries. Many PIVCs were idle, symptomatic, had substandard dressings, and were inserted in suboptimal anatomical sites. Despite international best practice guidelines, a large number of patients had PIVCs that were already failing or at risk of complications, including infection. A stronger focus is needed on compliance with PIVC insertion and management guidelines; better surveillance of PIVC sites; and improved assessment, decision-making, and documentation.

Acknowledgements

We are extremely grateful to colleagues from across the globe who committed their time and effort to this study (for full details of countries and team members see Appendix 1).