Original Research

 A bone island is a focus of normal cortical bone located within the medullary cavity. The vast majority of bone islands are small, measuring from 1 mm to 2 cm in size. They are found more frequently in adults than in children. The lesion can be virtually diagnosed on the basis of its characteristic clinical and imaging features. Differential diagnosis may be difficult when the lesion manifests itself uncharacteristically by being symptomatic, very large, and hot on bone scan.^1-4

The term giant bone island has been used to describe a large lesion1 that measures more than 2 cm in any dimension.⁵ Giant bone islands have been described only in adults,^1,5-15 and the longest bone island length reported is 10.5 cm.¹⁰ They are usually symptomatic and associated with increased radionuclide uptake on bone scintigraphy.¹⁴

The history and the clinical and imaging presentation of an even longer, symptomatic, and scintigraphically hot lesion in the tibial diaphysis of a 10-year-old boy is reported. The lesion further exhibited several atypical imaging features necessitating an open biopsy, which confirmed the diagnosis of a giant bone island. The pertinent differential diagnosis and the clinical and radiographic findings after 15-year follow-up are also presented and discussed. The patient provided written informed consent for print and electronic publication of this case report.

Case Report

A 10-year-old boy was admitted for surgical repair of an inguinal hernia. Physical examination revealed a painless but tender anterior bowing of the right tibial diaphysis. The patient was a healthy-appearing white male with normal vital signs, gait, and posture. His parents noticed a slight protuberance of the tibia at age 2.5 years. No medical advice was asked for the bone swelling after that time. After recovery from the inguinal hernia repair 3 weeks later, the bone lesion was thoroughly examined. Radiographs showed an oblong, homogenous region of dense sclerosis in the diaphysis of the right tibia. The lesion had relatively well-defined margins and was located in the medullary cavity. Speculations were not obvious in the periphery of the lesion, which exhibited a sharp circumscription (Figures 1A, 1B). A well-defined lytic area was evident at the distal part of the lesion (Figure 1B). There was no periosteal reaction. Blood and serum chemistries were within normal limits, including serum calcium, phosphorus, and alkaline phosphatase. A conventional 3-phase bone scintigraphy (300 MBq) with technetium-99m HDP (hydroxydiphosphonate) indicated increased uptake in the area of the lesion but no other skeletal abnormality (Figure 2). Computed tomography (CT) showed that the lesion was purely intramedullary and densely blastic. The lesion originated from the medial cortex, which was thickened (Figure 3A). The lesion extended to the anterolateral cortex, which was thinned and included a lytic area. In the distal part of the lesion, the anterolateral cortex was thickened, included lytic areas, and exhibited an anterior portion of cortical destruction (Figure 3B). The fatty marrow adjacent to the region of sclerosis appeared normal. There was no evidence of extraosseous soft-tissue changes. On both T1- and T2-weighted magnetic resonance imaging (MRI), the lesion exhibited low-signal intensity. The lesion measured 10.8×2.2×1 cm. It originated from the medial cortical bone of the tibia, blended into the medullary cavity, and extended anteriorly towards and through the anterior cortex. The area of cortical destruction was clearly evident on the axial MRI. The periosteum was displaced and eroded anteriorly by focal radiating bony streaks. No enhancement was seen after the intravenous administration of gadolinium-diethylenetriamine pentaacetic acid (Gd-DTPA) as a contrast medium. There were no extraosseous soft-tissue changes. In the distal part of the lesion, sagittal and axial MRI showed a 1.2×0.8×0.7-cm well-defined ovoid focus, with characteristics of cystic degeneration that exhibited intermediate-signal intensity on T1-weighted MRI (Figure 4) and high-signal intensity on T2-weighted MRI.

An open biopsy was performed. Macroscopically, a wedge of compact bone measuring 3×1.7×0.6 cm was taken. Microscopic examination showed a thinned periphery of lamellar (mature) bone with haversian canals and, beneath it, woven (immature) bone with long-surface processes projecting within adjacent cancellous bone (Figure 5A). The woven bone contained loose vascular fibrous tissue. No osteoclasts were noted, and the very few osteoblasts lining the bone trabeculae were small, single-layered, and flat (Figure 5B). There was no evidence of neoplastic cells. There was no abnormality of the periosteum and the surrounding soft tissues.

The histology was pathognomonic of a giant bone island. No additional surgical intervention was recommended.

The postoperative course was uncomplicated, and the patient was discharged 2 weeks later. An above-the-knee plaster was recommended for 3 months and a below-the-knee splint for an additional 2-month period. Full weight-bearing was allowed only after the postsurgical sixth month to prevent an impending fracture. The tibial bowing was tender to pressure or palpation, and the patient reported mild spontaneous pain during follow-up. Radiographs 1 year after surgery indicated that the bone area removed for biopsy was replaced by compact bone. MRI performed 4 years after surgery showed that the volume of the lesion in relation to the host bone was not changed.

At the last follow-up 15 years after surgery, the anterior tibial bowing was not changed (Figure 6A), but the patient additionally complained of skin irritation after intense training wearing boots during military service. The radiographic appearance of the lesion was also not changed, while the periphery of the lesion exhibited scarce radiating bony streaks with rounded contours (Figures 6B, 6C). The clinical symptoms and signs from wearing military boots completely subsided after a couple of weeks’ rest from daily army activities, but the mild spontaneous pain and the local tenderness over the tibial bowing persisted.

Discussion

Giant bone islands are more likely to be associated with clinical symptoms than the usual small-sized bone island. Some degree of pain was detected in 8 of 10 patients with a giant bone island presented in the literature, but it was induced by trauma in 3 of them.¹⁴

Radiographic appearance is among the distinguishing diagnostic features of a giant bone island. It appears as an ovoid, round, or oblong, homogenously dense, single or multiple focus of sclerosis within the medullary cavity; it is oriented along the long axis of the host bone, and it exhibits peripheral pseudopodia or radiating spicules producing the typical “thorny” or “paintbrush” appearance.^8,16,17 It does not exhibit cortical penetration and it is not associated with periosteal reaction.10

The CT findings include a sclerotic and hyperdense focus with spiculated margins extending into the adjacent cancellous bone. The lack of bone destruction and soft-tissue mass are also diagnostic.^3,7 MRI findings will reflect the low-signal intensity characteristics of cortical bone on all pulse sequences.¹⁸

Enostoses usually exhibit no activity on skeletal scintigraphy, while giant lesions generally show increased radiotracer uptake.^{5,9-11,14,19-27} The latter may result from the increased amount of bone turnover, which is seen more often with larger lesions because of active bone deposition and remodeling.^20,21,23,28 Histopathology of a giant bone island appears identical to the well-described pathologic appearance of smaller bone islands. The lesion is composed of compact lamellar bone and haversian systems, which blend with the adjacent spongiosa. The surrounding cancellous bone forms thorn-like trabeculae radiating from the lesion and merging with the cancellous bone.^1,4,5,8,28

The presumptive diagnosis of a bone island is based on the clinical findings, imaging features, and follow-up examinations. An asymptomatic, isolated, sclerotic bone lesion showing the typical features of a bone island on plain radiography, CT, and MRI, whatever its size, that is nonactive on bone scan may be easily diagnosed. However, a symptomatic patient with a hot lesion on scintigraphy should be carefully observed. In addition, larger lesions may raise the suspicion of a neoplasm, such as a sclerotic variant of osteosarcoma. In such cases, an open biopsy may be undertaken. No specific treatment is required after the diagnosis has been confirmed. There is no literature to suggest that, after adequate biopsy confirmation, excision or resection is necessary. Follow-up radiographic examination of the lesion should be suggested to monitor for any potential growth.^2,10,23

The first giant bone island appearing in a child is presented in this report. The lack of a causative factor leading to the anterior tibial bowing indicated that the bone deformity was caused primarily by the lesion. The present case is unusual for the appearance of several atypical features, some of which have not been previously described. Peripheral radiating spiculated margin was absent on the patient’s initial radiographs and CT imaging. MRI indicated only the presence of radiating bony streaks that displaced and eroded the periosteum on the anterior border of the lesion. The CT findings that the lesion likely originated or was in close proximity with the medial cortex of the tibia were also atypical. It has been previously reported that spinal lesions located immediately below the cortex tend to fuse with the endosteal surface, while similar features may also be seen in the appendicular enostoses.^4,29 Other CT findings, such as the thinning of the overlying anterolateral cortical bone, as well as the cortical thickening at the periphery of the lesion associated with areas of soft-tissue attenuation and anterior cortical destruction, have not been described even in the atypical features of a giant bone island. The lytic area resembling a nidus that was evident at the distal part of the lesion was more likely consistent with an area of resorption, which, although rare, has been described on giant lesions.^2,9,29 The substantial amount of woven bone transforming to lamellar bone that was evident in the present patient’s microscopic features is also an atypical finding, although it may be expected to some degree in scintigraphically hot, large lesions.²⁸ The clinical and imaging progress of the lesion supported the diagnosis of a giant bone island. The degree of the anterior tibial bowing and the volume of the lesion in relation to the host bone were not changed throughout the follow-up period, indicating that the growth of the lesion followed the growth of the normal bone.

The differential diagnosis of a giant bone island includes a variety of benign tumors and tumor-like lesions, as well as malignant bone lesions.^{2,4,23,28,30,31} In the patient presented in this report, the diagnosis of an atypical sclerotic presentation of a nonossifying fibroma or healing stage of this lesion could be consistent with some of the CT findings, including the eccentric origin from the cortex associated with medial cortical thickening, the anterolateral cortical thinning, and the soft-tissue attenuation of cortical areas. In addition, unifocal osteofibrous dysplasia may also present with a long intracortical diaphyseal lucency within an area of marked cortical sclerosis and cause a bowing deformity. Both diagnoses were excluded, since no fibrous stroma was evident on the histologic examination of the lesion. A large or giant long-bone osteoma would be associated with the outer cortical margin of bone but would not involve the intramedullary space. The scintigraphically increased uptake of radioisotope, as well as the CT and MRI findings, were not consistent with the diagnosis of osteoid osteoma, osteoblastoma, or osteomyelitis. Although most imaging findings were consistent with a benign lesion, and contrast-enhanced MRI showed no increased vascularity, anterior cortical disruption necessitated a bone biopsy to rule out any potential malignancy.

The histopathology in association with the clinical and imaging findings indicated the diagnosis of a giant bone island. The increased proportion of maturing woven bone over lamellar bone indicated an active remodeling lesion that could be related to the patient’s age, since the clinical and radiographic features of the lesion were not changed after 15-year follow-up.

Conclusion

This is the first giant bone island diagnosed in a patient before puberty. Its greatest length was 10.8 cm, which is the longest reported in the literature. The imaging appearance included several atypical features that are very rare or have not been reported. Microscopic features indicated less mature lamellar bone and a prominent proportion of maturing woven bone. The clinical and the radiographic appearance of the lesion were not changed after 15-year follow-up.

 A bone island is a focus of normal cortical bone located within the medullary cavity. The vast majority of bone islands are small, measuring from 1 mm to 2 cm in size. They are found more frequently in adults than in children. The lesion can be virtually diagnosed on the basis of its characteristic clinical and imaging features. Differential diagnosis may be difficult when the lesion manifests itself uncharacteristically by being symptomatic, very large, and hot on bone scan.^1-4

The term giant bone island has been used to describe a large lesion1 that measures more than 2 cm in any dimension.⁵ Giant bone islands have been described only in adults,^1,5-15 and the longest bone island length reported is 10.5 cm.¹⁰ They are usually symptomatic and associated with increased radionuclide uptake on bone scintigraphy.¹⁴

The history and the clinical and imaging presentation of an even longer, symptomatic, and scintigraphically hot lesion in the tibial diaphysis of a 10-year-old boy is reported. The lesion further exhibited several atypical imaging features necessitating an open biopsy, which confirmed the diagnosis of a giant bone island. The pertinent differential diagnosis and the clinical and radiographic findings after 15-year follow-up are also presented and discussed. The patient provided written informed consent for print and electronic publication of this case report.

Case Report

A 10-year-old boy was admitted for surgical repair of an inguinal hernia. Physical examination revealed a painless but tender anterior bowing of the right tibial diaphysis. The patient was a healthy-appearing white male with normal vital signs, gait, and posture. His parents noticed a slight protuberance of the tibia at age 2.5 years. No medical advice was asked for the bone swelling after that time. After recovery from the inguinal hernia repair 3 weeks later, the bone lesion was thoroughly examined. Radiographs showed an oblong, homogenous region of dense sclerosis in the diaphysis of the right tibia. The lesion had relatively well-defined margins and was located in the medullary cavity. Speculations were not obvious in the periphery of the lesion, which exhibited a sharp circumscription (Figures 1A, 1B). A well-defined lytic area was evident at the distal part of the lesion (Figure 1B). There was no periosteal reaction. Blood and serum chemistries were within normal limits, including serum calcium, phosphorus, and alkaline phosphatase. A conventional 3-phase bone scintigraphy (300 MBq) with technetium-99m HDP (hydroxydiphosphonate) indicated increased uptake in the area of the lesion but no other skeletal abnormality (Figure 2). Computed tomography (CT) showed that the lesion was purely intramedullary and densely blastic. The lesion originated from the medial cortex, which was thickened (Figure 3A). The lesion extended to the anterolateral cortex, which was thinned and included a lytic area. In the distal part of the lesion, the anterolateral cortex was thickened, included lytic areas, and exhibited an anterior portion of cortical destruction (Figure 3B). The fatty marrow adjacent to the region of sclerosis appeared normal. There was no evidence of extraosseous soft-tissue changes. On both T1- and T2-weighted magnetic resonance imaging (MRI), the lesion exhibited low-signal intensity. The lesion measured 10.8×2.2×1 cm. It originated from the medial cortical bone of the tibia, blended into the medullary cavity, and extended anteriorly towards and through the anterior cortex. The area of cortical destruction was clearly evident on the axial MRI. The periosteum was displaced and eroded anteriorly by focal radiating bony streaks. No enhancement was seen after the intravenous administration of gadolinium-diethylenetriamine pentaacetic acid (Gd-DTPA) as a contrast medium. There were no extraosseous soft-tissue changes. In the distal part of the lesion, sagittal and axial MRI showed a 1.2×0.8×0.7-cm well-defined ovoid focus, with characteristics of cystic degeneration that exhibited intermediate-signal intensity on T1-weighted MRI (Figure 4) and high-signal intensity on T2-weighted MRI.

An open biopsy was performed. Macroscopically, a wedge of compact bone measuring 3×1.7×0.6 cm was taken. Microscopic examination showed a thinned periphery of lamellar (mature) bone with haversian canals and, beneath it, woven (immature) bone with long-surface processes projecting within adjacent cancellous bone (Figure 5A). The woven bone contained loose vascular fibrous tissue. No osteoclasts were noted, and the very few osteoblasts lining the bone trabeculae were small, single-layered, and flat (Figure 5B). There was no evidence of neoplastic cells. There was no abnormality of the periosteum and the surrounding soft tissues.

The histology was pathognomonic of a giant bone island. No additional surgical intervention was recommended.

The postoperative course was uncomplicated, and the patient was discharged 2 weeks later. An above-the-knee plaster was recommended for 3 months and a below-the-knee splint for an additional 2-month period. Full weight-bearing was allowed only after the postsurgical sixth month to prevent an impending fracture. The tibial bowing was tender to pressure or palpation, and the patient reported mild spontaneous pain during follow-up. Radiographs 1 year after surgery indicated that the bone area removed for biopsy was replaced by compact bone. MRI performed 4 years after surgery showed that the volume of the lesion in relation to the host bone was not changed.

At the last follow-up 15 years after surgery, the anterior tibial bowing was not changed (Figure 6A), but the patient additionally complained of skin irritation after intense training wearing boots during military service. The radiographic appearance of the lesion was also not changed, while the periphery of the lesion exhibited scarce radiating bony streaks with rounded contours (Figures 6B, 6C). The clinical symptoms and signs from wearing military boots completely subsided after a couple of weeks’ rest from daily army activities, but the mild spontaneous pain and the local tenderness over the tibial bowing persisted.

Discussion

Giant bone islands are more likely to be associated with clinical symptoms than the usual small-sized bone island. Some degree of pain was detected in 8 of 10 patients with a giant bone island presented in the literature, but it was induced by trauma in 3 of them.¹⁴

Radiographic appearance is among the distinguishing diagnostic features of a giant bone island. It appears as an ovoid, round, or oblong, homogenously dense, single or multiple focus of sclerosis within the medullary cavity; it is oriented along the long axis of the host bone, and it exhibits peripheral pseudopodia or radiating spicules producing the typical “thorny” or “paintbrush” appearance.^8,16,17 It does not exhibit cortical penetration and it is not associated with periosteal reaction.10

The CT findings include a sclerotic and hyperdense focus with spiculated margins extending into the adjacent cancellous bone. The lack of bone destruction and soft-tissue mass are also diagnostic.^3,7 MRI findings will reflect the low-signal intensity characteristics of cortical bone on all pulse sequences.¹⁸

Enostoses usually exhibit no activity on skeletal scintigraphy, while giant lesions generally show increased radiotracer uptake.^{5,9-11,14,19-27} The latter may result from the increased amount of bone turnover, which is seen more often with larger lesions because of active bone deposition and remodeling.^20,21,23,28 Histopathology of a giant bone island appears identical to the well-described pathologic appearance of smaller bone islands. The lesion is composed of compact lamellar bone and haversian systems, which blend with the adjacent spongiosa. The surrounding cancellous bone forms thorn-like trabeculae radiating from the lesion and merging with the cancellous bone.^1,4,5,8,28

The presumptive diagnosis of a bone island is based on the clinical findings, imaging features, and follow-up examinations. An asymptomatic, isolated, sclerotic bone lesion showing the typical features of a bone island on plain radiography, CT, and MRI, whatever its size, that is nonactive on bone scan may be easily diagnosed. However, a symptomatic patient with a hot lesion on scintigraphy should be carefully observed. In addition, larger lesions may raise the suspicion of a neoplasm, such as a sclerotic variant of osteosarcoma. In such cases, an open biopsy may be undertaken. No specific treatment is required after the diagnosis has been confirmed. There is no literature to suggest that, after adequate biopsy confirmation, excision or resection is necessary. Follow-up radiographic examination of the lesion should be suggested to monitor for any potential growth.^2,10,23

The first giant bone island appearing in a child is presented in this report. The lack of a causative factor leading to the anterior tibial bowing indicated that the bone deformity was caused primarily by the lesion. The present case is unusual for the appearance of several atypical features, some of which have not been previously described. Peripheral radiating spiculated margin was absent on the patient’s initial radiographs and CT imaging. MRI indicated only the presence of radiating bony streaks that displaced and eroded the periosteum on the anterior border of the lesion. The CT findings that the lesion likely originated or was in close proximity with the medial cortex of the tibia were also atypical. It has been previously reported that spinal lesions located immediately below the cortex tend to fuse with the endosteal surface, while similar features may also be seen in the appendicular enostoses.^4,29 Other CT findings, such as the thinning of the overlying anterolateral cortical bone, as well as the cortical thickening at the periphery of the lesion associated with areas of soft-tissue attenuation and anterior cortical destruction, have not been described even in the atypical features of a giant bone island. The lytic area resembling a nidus that was evident at the distal part of the lesion was more likely consistent with an area of resorption, which, although rare, has been described on giant lesions.^2,9,29 The substantial amount of woven bone transforming to lamellar bone that was evident in the present patient’s microscopic features is also an atypical finding, although it may be expected to some degree in scintigraphically hot, large lesions.²⁸ The clinical and imaging progress of the lesion supported the diagnosis of a giant bone island. The degree of the anterior tibial bowing and the volume of the lesion in relation to the host bone were not changed throughout the follow-up period, indicating that the growth of the lesion followed the growth of the normal bone.

The differential diagnosis of a giant bone island includes a variety of benign tumors and tumor-like lesions, as well as malignant bone lesions.^{2,4,23,28,30,31} In the patient presented in this report, the diagnosis of an atypical sclerotic presentation of a nonossifying fibroma or healing stage of this lesion could be consistent with some of the CT findings, including the eccentric origin from the cortex associated with medial cortical thickening, the anterolateral cortical thinning, and the soft-tissue attenuation of cortical areas. In addition, unifocal osteofibrous dysplasia may also present with a long intracortical diaphyseal lucency within an area of marked cortical sclerosis and cause a bowing deformity. Both diagnoses were excluded, since no fibrous stroma was evident on the histologic examination of the lesion. A large or giant long-bone osteoma would be associated with the outer cortical margin of bone but would not involve the intramedullary space. The scintigraphically increased uptake of radioisotope, as well as the CT and MRI findings, were not consistent with the diagnosis of osteoid osteoma, osteoblastoma, or osteomyelitis. Although most imaging findings were consistent with a benign lesion, and contrast-enhanced MRI showed no increased vascularity, anterior cortical disruption necessitated a bone biopsy to rule out any potential malignancy.

The histopathology in association with the clinical and imaging findings indicated the diagnosis of a giant bone island. The increased proportion of maturing woven bone over lamellar bone indicated an active remodeling lesion that could be related to the patient’s age, since the clinical and radiographic features of the lesion were not changed after 15-year follow-up.

Conclusion

This is the first giant bone island diagnosed in a patient before puberty. Its greatest length was 10.8 cm, which is the longest reported in the literature. The imaging appearance included several atypical features that are very rare or have not been reported. Microscopic features indicated less mature lamellar bone and a prominent proportion of maturing woven bone. The clinical and the radiographic appearance of the lesion were not changed after 15-year follow-up.

 A bone island is a focus of normal cortical bone located within the medullary cavity. The vast majority of bone islands are small, measuring from 1 mm to 2 cm in size. They are found more frequently in adults than in children. The lesion can be virtually diagnosed on the basis of its characteristic clinical and imaging features. Differential diagnosis may be difficult when the lesion manifests itself uncharacteristically by being symptomatic, very large, and hot on bone scan.^1-4

The term giant bone island has been used to describe a large lesion1 that measures more than 2 cm in any dimension.⁵ Giant bone islands have been described only in adults,^1,5-15 and the longest bone island length reported is 10.5 cm.¹⁰ They are usually symptomatic and associated with increased radionuclide uptake on bone scintigraphy.¹⁴

The history and the clinical and imaging presentation of an even longer, symptomatic, and scintigraphically hot lesion in the tibial diaphysis of a 10-year-old boy is reported. The lesion further exhibited several atypical imaging features necessitating an open biopsy, which confirmed the diagnosis of a giant bone island. The pertinent differential diagnosis and the clinical and radiographic findings after 15-year follow-up are also presented and discussed. The patient provided written informed consent for print and electronic publication of this case report.

Case Report

A 10-year-old boy was admitted for surgical repair of an inguinal hernia. Physical examination revealed a painless but tender anterior bowing of the right tibial diaphysis. The patient was a healthy-appearing white male with normal vital signs, gait, and posture. His parents noticed a slight protuberance of the tibia at age 2.5 years. No medical advice was asked for the bone swelling after that time. After recovery from the inguinal hernia repair 3 weeks later, the bone lesion was thoroughly examined. Radiographs showed an oblong, homogenous region of dense sclerosis in the diaphysis of the right tibia. The lesion had relatively well-defined margins and was located in the medullary cavity. Speculations were not obvious in the periphery of the lesion, which exhibited a sharp circumscription (Figures 1A, 1B). A well-defined lytic area was evident at the distal part of the lesion (Figure 1B). There was no periosteal reaction. Blood and serum chemistries were within normal limits, including serum calcium, phosphorus, and alkaline phosphatase. A conventional 3-phase bone scintigraphy (300 MBq) with technetium-99m HDP (hydroxydiphosphonate) indicated increased uptake in the area of the lesion but no other skeletal abnormality (Figure 2). Computed tomography (CT) showed that the lesion was purely intramedullary and densely blastic. The lesion originated from the medial cortex, which was thickened (Figure 3A). The lesion extended to the anterolateral cortex, which was thinned and included a lytic area. In the distal part of the lesion, the anterolateral cortex was thickened, included lytic areas, and exhibited an anterior portion of cortical destruction (Figure 3B). The fatty marrow adjacent to the region of sclerosis appeared normal. There was no evidence of extraosseous soft-tissue changes. On both T1- and T2-weighted magnetic resonance imaging (MRI), the lesion exhibited low-signal intensity. The lesion measured 10.8×2.2×1 cm. It originated from the medial cortical bone of the tibia, blended into the medullary cavity, and extended anteriorly towards and through the anterior cortex. The area of cortical destruction was clearly evident on the axial MRI. The periosteum was displaced and eroded anteriorly by focal radiating bony streaks. No enhancement was seen after the intravenous administration of gadolinium-diethylenetriamine pentaacetic acid (Gd-DTPA) as a contrast medium. There were no extraosseous soft-tissue changes. In the distal part of the lesion, sagittal and axial MRI showed a 1.2×0.8×0.7-cm well-defined ovoid focus, with characteristics of cystic degeneration that exhibited intermediate-signal intensity on T1-weighted MRI (Figure 4) and high-signal intensity on T2-weighted MRI.

An open biopsy was performed. Macroscopically, a wedge of compact bone measuring 3×1.7×0.6 cm was taken. Microscopic examination showed a thinned periphery of lamellar (mature) bone with haversian canals and, beneath it, woven (immature) bone with long-surface processes projecting within adjacent cancellous bone (Figure 5A). The woven bone contained loose vascular fibrous tissue. No osteoclasts were noted, and the very few osteoblasts lining the bone trabeculae were small, single-layered, and flat (Figure 5B). There was no evidence of neoplastic cells. There was no abnormality of the periosteum and the surrounding soft tissues.

The histology was pathognomonic of a giant bone island. No additional surgical intervention was recommended.

The postoperative course was uncomplicated, and the patient was discharged 2 weeks later. An above-the-knee plaster was recommended for 3 months and a below-the-knee splint for an additional 2-month period. Full weight-bearing was allowed only after the postsurgical sixth month to prevent an impending fracture. The tibial bowing was tender to pressure or palpation, and the patient reported mild spontaneous pain during follow-up. Radiographs 1 year after surgery indicated that the bone area removed for biopsy was replaced by compact bone. MRI performed 4 years after surgery showed that the volume of the lesion in relation to the host bone was not changed.

At the last follow-up 15 years after surgery, the anterior tibial bowing was not changed (Figure 6A), but the patient additionally complained of skin irritation after intense training wearing boots during military service. The radiographic appearance of the lesion was also not changed, while the periphery of the lesion exhibited scarce radiating bony streaks with rounded contours (Figures 6B, 6C). The clinical symptoms and signs from wearing military boots completely subsided after a couple of weeks’ rest from daily army activities, but the mild spontaneous pain and the local tenderness over the tibial bowing persisted.

Discussion

Giant bone islands are more likely to be associated with clinical symptoms than the usual small-sized bone island. Some degree of pain was detected in 8 of 10 patients with a giant bone island presented in the literature, but it was induced by trauma in 3 of them.¹⁴

Radiographic appearance is among the distinguishing diagnostic features of a giant bone island. It appears as an ovoid, round, or oblong, homogenously dense, single or multiple focus of sclerosis within the medullary cavity; it is oriented along the long axis of the host bone, and it exhibits peripheral pseudopodia or radiating spicules producing the typical “thorny” or “paintbrush” appearance.^8,16,17 It does not exhibit cortical penetration and it is not associated with periosteal reaction.10

The CT findings include a sclerotic and hyperdense focus with spiculated margins extending into the adjacent cancellous bone. The lack of bone destruction and soft-tissue mass are also diagnostic.^3,7 MRI findings will reflect the low-signal intensity characteristics of cortical bone on all pulse sequences.¹⁸

Enostoses usually exhibit no activity on skeletal scintigraphy, while giant lesions generally show increased radiotracer uptake.^{5,9-11,14,19-27} The latter may result from the increased amount of bone turnover, which is seen more often with larger lesions because of active bone deposition and remodeling.^20,21,23,28 Histopathology of a giant bone island appears identical to the well-described pathologic appearance of smaller bone islands. The lesion is composed of compact lamellar bone and haversian systems, which blend with the adjacent spongiosa. The surrounding cancellous bone forms thorn-like trabeculae radiating from the lesion and merging with the cancellous bone.^1,4,5,8,28

The presumptive diagnosis of a bone island is based on the clinical findings, imaging features, and follow-up examinations. An asymptomatic, isolated, sclerotic bone lesion showing the typical features of a bone island on plain radiography, CT, and MRI, whatever its size, that is nonactive on bone scan may be easily diagnosed. However, a symptomatic patient with a hot lesion on scintigraphy should be carefully observed. In addition, larger lesions may raise the suspicion of a neoplasm, such as a sclerotic variant of osteosarcoma. In such cases, an open biopsy may be undertaken. No specific treatment is required after the diagnosis has been confirmed. There is no literature to suggest that, after adequate biopsy confirmation, excision or resection is necessary. Follow-up radiographic examination of the lesion should be suggested to monitor for any potential growth.^2,10,23

The first giant bone island appearing in a child is presented in this report. The lack of a causative factor leading to the anterior tibial bowing indicated that the bone deformity was caused primarily by the lesion. The present case is unusual for the appearance of several atypical features, some of which have not been previously described. Peripheral radiating spiculated margin was absent on the patient’s initial radiographs and CT imaging. MRI indicated only the presence of radiating bony streaks that displaced and eroded the periosteum on the anterior border of the lesion. The CT findings that the lesion likely originated or was in close proximity with the medial cortex of the tibia were also atypical. It has been previously reported that spinal lesions located immediately below the cortex tend to fuse with the endosteal surface, while similar features may also be seen in the appendicular enostoses.^4,29 Other CT findings, such as the thinning of the overlying anterolateral cortical bone, as well as the cortical thickening at the periphery of the lesion associated with areas of soft-tissue attenuation and anterior cortical destruction, have not been described even in the atypical features of a giant bone island. The lytic area resembling a nidus that was evident at the distal part of the lesion was more likely consistent with an area of resorption, which, although rare, has been described on giant lesions.^2,9,29 The substantial amount of woven bone transforming to lamellar bone that was evident in the present patient’s microscopic features is also an atypical finding, although it may be expected to some degree in scintigraphically hot, large lesions.²⁸ The clinical and imaging progress of the lesion supported the diagnosis of a giant bone island. The degree of the anterior tibial bowing and the volume of the lesion in relation to the host bone were not changed throughout the follow-up period, indicating that the growth of the lesion followed the growth of the normal bone.

The differential diagnosis of a giant bone island includes a variety of benign tumors and tumor-like lesions, as well as malignant bone lesions.^{2,4,23,28,30,31} In the patient presented in this report, the diagnosis of an atypical sclerotic presentation of a nonossifying fibroma or healing stage of this lesion could be consistent with some of the CT findings, including the eccentric origin from the cortex associated with medial cortical thickening, the anterolateral cortical thinning, and the soft-tissue attenuation of cortical areas. In addition, unifocal osteofibrous dysplasia may also present with a long intracortical diaphyseal lucency within an area of marked cortical sclerosis and cause a bowing deformity. Both diagnoses were excluded, since no fibrous stroma was evident on the histologic examination of the lesion. A large or giant long-bone osteoma would be associated with the outer cortical margin of bone but would not involve the intramedullary space. The scintigraphically increased uptake of radioisotope, as well as the CT and MRI findings, were not consistent with the diagnosis of osteoid osteoma, osteoblastoma, or osteomyelitis. Although most imaging findings were consistent with a benign lesion, and contrast-enhanced MRI showed no increased vascularity, anterior cortical disruption necessitated a bone biopsy to rule out any potential malignancy.

The histopathology in association with the clinical and imaging findings indicated the diagnosis of a giant bone island. The increased proportion of maturing woven bone over lamellar bone indicated an active remodeling lesion that could be related to the patient’s age, since the clinical and radiographic features of the lesion were not changed after 15-year follow-up.

Conclusion

This is the first giant bone island diagnosed in a patient before puberty. Its greatest length was 10.8 cm, which is the longest reported in the literature. The imaging appearance included several atypical features that are very rare or have not been reported. Microscopic features indicated less mature lamellar bone and a prominent proportion of maturing woven bone. The clinical and the radiographic appearance of the lesion were not changed after 15-year follow-up.

Streptococcus pneumoniae (S pneumoniae), also known as pneumococcus, is a successful human pathogen with significant clinical impact that causes pneumonia and invasive infections, including bacteremia and meningitis.¹ In the preantibiotic era, nearly 80% of patients with bacteremic pneumococcal pneumonia died.² The introduction of sulfas and penicillin in the mid-20th century, subsequent refinements in antibiotic chemotherapy, and improvements in supportive care rendered pneumococcal disease readily treatable, notwithstanding the threat of antibiotic-resistant pneumococcus.³ Despite the availability of effective antibiotic therapy against S pneumoniae, pneumococcal disease remains a significant cause of morbidity and mortality among people with increased susceptibility, such as older adults and those living with chronic illness or immunosuppressive conditions. In developed countries like the U.S., where a growing portion of the population is vulnerable to S pneumoniae by virtue of their advanced age and underlying medical conditions, pneumococcal disease is still an important public health concern.⁴

Penumococcal Vaccines: A Long Time Coming

Vaccination against S pneumoniae has proven an efficacious strategy to reduce the morbidity and mortality associated with this pathogen.⁵ The original efforts to develop a pneumococcal vaccine culminated in 1945 with a vaccine containing pneumococcal capsular polysaccharides, which elicited a protective immune response among U.S. soldiers (Table 1).Subsequent investigations determined that the protective response was specific to pneumococcal disease caused by the 4 pneumococcal capsular serotypes included in the vaccine, that the carrier rate of pneumococcus with the vaccine serotypes decreased by about 50%, and that the incidence of pneumonia from the vaccine serotypes was reduced even in nonimmunized soldiers.⁶ These early observations remain relevant to our contemporary understanding of the impact of pneumococcal vaccination: Protection is limited to serotypes included in the vaccine; the vaccine reduces colonization; and the vaccine leads to herd immunity—the protection of unvaccinated subjects.

Despite this achievement, the use of vaccines as a strategy to combat pneumococcal disease was upstaged in the subsequent decades by the success of antibiotics. Renewed interest in pneumococcal vaccines resulted from the efforts of Robert Austrian, MD, who astutely observed that “highly effective antimicrobial drugs must be supplemented by other measures, both prophylactic and therapeutic, if the significant mortality resulting still from pneumococcal infection is to be reduced.”⁷

After initial studies carried out in South African gold miners, the FDA approved a pneumococcal polysaccharide vaccine (PPSV) against 23 of about 90 circulating serotypes.⁸ The CDC and the Advisory Committee on Immunization Practices (ACIP) initially recommended PPSV-23 for persons perceived to be at high risk for pneumococcal disease, including those with chronic diseases, immunocompromising conditions, and older adults.⁹

Three decades later, the analysis of a large body of available evidence demonstrated the protective effects of PPSV-23 against invasive pneumococcal disease caused by pneumococcal types included in the vaccine, especially bacteremic pneumonia (about 75% reduction, odds ratio [OR] 0.26, 95% confidence interval [CI] 0.14-0.45).¹⁰ A perceived shortcoming of PPSV-23, which some experts dispute, is the lack of definite protection against nonbacteremic pneumococcal pneumonia.¹¹ Studies of nursing home residents demonstrated a 50% reduction in the incidence of both pneumococcal pneumonia and all-cause pneumonia, suggesting that PPSV-23 offers protection against noninvasive pneumococcal pneumonia in specific populations.¹² Additional potential limitations of PPSV-23 include reduced benefit in patients aged > 65 years and waning of protection over time.¹³

Related: Identification and Management of Middle East Respiratory Syndrome

Protein-Conjugate Vaccines

Clearly, the most important limitation of PPSV-23, inherent to all capsular polysaccharide vaccines, is that it does not elicit an immune response in children aged < 2 years. The successful development of a vaccine against Haemophilus influenzae type b (Hib) gave rise to a new generation of pneumococcal vaccines.¹⁴ Specifically, the Hib vaccine covalently binds, or conjugates, the capsule polysaccharide to an antigenic protein, leading to effective T-cell–mediated antibody production in infants and toddlers.

In 2000, children received the first iteration of a protein-conjugate vaccine containing the 7 most relevant pneumococcal serotypes (PCV-7).¹⁵ The effects of PCV-7 on pneumococcal disease have been extraordinary, practically eliminating infections caused by the pneumococcal serotypes included in the vaccine. Immunizing children with PCV-7 also ushered in a fundamental public health benefit for adults aged > 65 years: a reduction of nearly 90% in the incidence of pneumococcal infections caused by serotypes included in the vaccine. By eliciting mucosal immunity, which leads to decreased nasal carriage of the covered pneumococcal strains among children, PVC-7 generates herd immunity, leading to reductions in transmission, colonization, and infection with vaccine serotypes among adults.¹⁶

In 2010, a conjugate vaccine containing 13 serotypes (PCV-13) replaced PCV-7 administration for children.^17,18 The PCV-13 is expected to protect children and the herd from disease caused by 6 additional pneumococcal serotypes, including those that surged as replacement strains, filling the ecologic niche created by PCV-7, such as the epidemiologically relevant serotype 19A.^19,20

Immune-Compromised Adults

One of the shortcomings of PPSV-23 is its lack of efficacy in patients with advanced HIV, a group with an exquisite vulnerability to pneumococcal disease, as demonstrated by a randomized controlled trial of PPSV-23 in African patients with advanced, untreated HIV infection.²¹ Similarly, there is concern that patients with lymphoma, leukemia, multiple myeloma, and Hodgkin disease, also at high risk of pneumococcal infection, do not mount sufficient immune responses to polysaccharide antigens or that these responses are adversely affected by chemotherapy or immune suppressing medications. In these populations, a conjugate vaccine (PCV-7 or PCV-13) may elicit a more robust and durable immune response than a polysaccharide vaccine (PPSV-23). A randomized placebo-controlled trial demonstrated the efficacy of PCV-7 in protecting patients with advanced HIV against pneumococcal disease in Africa.²² Based in part on this observation, the ACIP now recommends the use of PCV-13 in patients with HIV and other immunosuppressing conditions, including chronic renal failure.²³ A direct comparison of the relative protection of PPSV-23 vs PCV-7 in this population has not been performed.

Pneumococcal Vaccines in the Elderly

Although PPSV-23 was widely adopted in the U.S. with the intention to protect adults aged > 65 years from pneumococcal infection, this vaccine did not achieve global appeal. In the Netherlands, for instance, PPSV-23 had very low penetration and was never endorsed by public health authorities. Concerns have been issued about the decreased immunogenicity of a polysaccharide vaccine in older adults, invoking the concept of immune senescence, a term that describes a diminished capacity to mount robust immune responses due to aging.

Protein-conjugated antigens, on the other hand, can elicit a better initial immune response than a polysaccharide vaccine can in an older adult. Whether the effect is sustained and translates to better clinical outcomes remains unknown. The adoption of PCV-13 and the continued role of PPSV-23 in adults aged > 65 years have been examined carefully by experts.^24,25

Dutch investigators organized the Community Acquired Pneumonia Immunization Trial in Adults (CAPITA), randomizing about 85,000 adults aged > 65 years toreceive eitherPCV-13 or a placebo.²⁶ After 3 years of observation, the occurrence of invasive pneumococcal disease (caused by the vaccine serotypes) decreased by 85% in participants who received the vaccine compared with those who received a placebo. Additionally, the incidence of pneumococcal pneumonia (caused by the vaccine serotypes) decreased by 45% (Table 2). The findings of this trial largely inform the recommendation issued by APIC in October 2014 to administer PCV-13 vaccine to adults aged > 65 years.²⁷

Polysaccharide Vaccine in Adults

A crucial limitation of the CAPITA study is that it does not provide a head-to-head comparison with PPSV-23. Similarly, the recommendation to use PCV-13 in patients with immune system compromising conditions did not arise from a direct comparison between PCV-13 and PPSV-23. In both groups of patients, ACIP continues to recommend PPSV-23 to extend protection to the 10 additional pneumococcal serotypes not included in PCV-13.²⁸ Additionally, ACIP maintains its long-standing recommendation to administer PPSV-23 to adults aged < 65 years who have diabetes mellitus, or chronic liver, lung, or heart disease, or those who are tobacco smokers or misuse alcohol.

Notably, the order in which the vaccines are administered may influence their effectiveness. In adults aged 50 to 64 years, initial vaccination with PCV-13 followed by PPSV-23 lead to a better antipneumococcal immune response.^14,29 Specifically, PCV-13 enhanced the response to subsequent administration of PPSV-23, whereas an initial PPSV-23 vaccine resulted in a decreased response to subsequent administration of PCV-13. After a 1-year interval between vaccines, immune titers to a second vaccination were not inferior.³⁰ All the above considerations have shaped the ACIP recommendations for the administration of pneumococcal vaccines.^27,31 Pneumococcal vaccination was a straightforward exercise when PPSV-23 was the only vaccine available for adults. The advent of PCV-13, however, renders pneumococcal immunization of adults more complicated (Figure 1).

Improving Vaccination Rates

The complex recommendations to vaccinate adults against pneumococcus may reveal obstacles to effective pneumococcal vaccination programs.³² From the perspective of health care institutions, the adoption of the new pneumococcal vaccine implies an added cost. A pneumococcal vaccination strategy that incorporates PCV-13 may be cost-effective, at least under certain parameters.^33-35 There remains, however, the issue of affordability and the opportunity cost—the need to decrease funding for other health care programs to accommodate an increased budget for pneumococcal vaccination. Logistically, maintaining a reliable supply of vaccines to meet the demand of practitioners at various sites requires careful planning. Consequently, tracking vaccinated and unvaccinated patients and carefully coordinating among clinical providers, nurses, and pharmacists are essential.

From the perspective of clinical providers, an additional pneumococcal vaccine complicating the vaccination schedule for adults represents an increased burden. Providers will need information to reach their own conclusions regarding the rationale behind the development and use of pneumococcal conjugate vaccines, the existing and evolving recommendations from public health authorities, and the strengths and limitations of the evidence supporting the use of pneumococcal vaccines. Otherwise, providers may find it difficult to incorporate new data and guidelines supporting pneumococcal vaccination into their decision-making (Boxes 1-4). An additional and formidable challenge is to carve out time during an already busy clinical encounter to discuss pneumococcal vaccines and other immunizations.³⁶

Related: The Importance of an Antimicrobial Stewardship Program

Similarly, older adults or patients with chronic health conditions may not recognize the important role that vaccines can play in their health maintenance and are likely to prioritize other issues during medical visits. It is not obvious for patients that multiple vaccines may be necessary to prevent pneumococcal disease. Furthermore, many patients, not unreasonably, may assume that their yearly influenza vaccine is sufficiently protective against pneumonia. Therefore, patients need to be educated about the rationale behind pneumococcal conjugate vaccination. Ultimately, access to immunization—the opportunity for patients to have an encounter with their providers and with the health care system that results in the administration of an appropriate vaccine—will determine whether goals for pneumococcal vaccination are achieved.

The evolving landscape for the implementation of pneumococcal vaccines creates the need to develop, implement, and refine organizational changes to adhere to the new guidelines for the use of PCV-13 and PPSV-23 vaccines. These interventions, if effective, may help improve pneumococcal vaccination coverage among adults (Table 3).

Harnessing the Power of the EMR

Interventions to improve the adherence to pneumococcal vaccination guidelines begin by identifying persons eligible for vaccination based on their age, their vaccination status (ie, persons previously unvaccinated or due for vaccination according to the recommended schedule), or the presence of medical conditions conferring high risk for pneumococcal disease. This, in turn, depends on adequate documentation of patients’ underlying medical diagnosis, as well as up-to-date records of vaccine administration to patients.

Health care systems possessing a mature and sophisticated electronic medical record (EMR), such as the VHA Computerized Patient Record System (CPRS), are in a good position to wield such information to plan, implement, and assure the quality of activities designed to improve pneumococcal vaccination rates. An analysis of the proportion of veterans in VISN 10 who received pneumococcal vaccination revealed that even with the advantages of a robust EMR and a highly developed infrastructure devoted to primary care, pneumococcal vaccine coverage remains below the 60% target goal, although well above national averages (Figure 2).^37,38

Standing Orders

Standing orders make it possible for nurses and pharmacists to administer vaccines according to a preestablished protocol without a physician’s direct evaluation of each patient. Standing orders are a versatile intervention, with a record of effective implementation in both outpatient and inpatient facilities, in acute-care and long-term care facilities, and in most instances where patients interact with the health care system. Based on strong scientific evidence, ACIP recommends the adoption of standing-order programs to improve pneumococcal vaccination rates among adults.³⁹ Indeed, standing-order programs may prove a very effective intervention to fulfill the recommendation to administer PCV-13 to adults aged > 65 years.

The Virtual Vaccination Clinic

Unfortunately, with 2 vaccines that have to be administered at different times to various groups of patients at risk, the current state of pneumococcal vaccination may be too complex to be readily reduced to a comprehensible set of standing orders. An innovative way to realize the benefit of standing orders is to target high-risk groups for pneumococcal disease who are eligible for vaccination by selecting them using the EMR and entering standing orders tailored to their specific vaccination needs. The selection of patients according to comorbidities and vaccination status and the determination of the appropriate pneumococcal vaccine takes place in the context of a “virtual” vaccination clinic.⁴⁰

Enhancing Vaccinations


Further improvement in pneumococcal vaccination rates are likely to result from interventions that increase the demand for vaccines among patients and practitioners. Efforts to disseminate information and provide advice regarding pneumococcal vaccination are likely to result in patients seeking and clinicians offering the appropriate vaccine. Similarly, interventions to enhance the supply of vaccines at the point of care may reduce barriers that patients might encounter when attempting to receive vaccinations.

Another set of system-based interventions that can assist clinicians in making timely and appropriate vaccination decisions are EMR reminders, especially those targeting patients at high risk for pneumococcal disease because of underlying illnesses.⁴¹ Previous experience with pneumococcal vaccination in patients aged > 65 years indicates that EMR reminders facilitate improvements in vaccination. The involvement of a panel manager who coordinated with the primary care provider and contacted patients directly augmented the effect of the EMR reminder by 25%.⁴²

Related: Venous Thromboembolism Prophylaxis in Acutely Ill Veterans With Respiratory Disease

Patient reminder and recall systems also demonstrated effectiveness in improving immunization rates.⁴³ In certain groups, notification of patients has been achieved through electronic methods, such as short text messaging or e-mail.⁴⁴ Determining which interventions within a bundle are essential may be impossible, because the various interventions reinforce one another, and the likelihood of patients benefitting from at least one of the activities increases when multiple interventions are administered together. Therefore, the Task Force on Community Preventive Services supports combining provider reminder systems with education and other measures that encourage use of vaccines in patients and providers.⁴⁵

Box answer key: 1: A; 2: A; 3: B; 4: A.

Conclusion

The increasing role of vaccines in the health maintenance of adults represents a change in paradigm for primary care and specialty providers. Physicians must assess the value and limitations of vaccines and find time to discuss immunizations with their adult patients. Health care systems can increase opportunities for vaccination and facilitate encounters that result in vaccination by educating patients and health care personnel and through the innovative use of reminders and standing orders in the EMR. Undertaking these activities may limit the burden of pneumococcal disease, an important cause of morbidity and mortality in adults that is preventable through immunization.

Acknowledgements
This work is dedicated to the memory of John M. Rieger, Information Technology Specialist at the Cleveland VAMC and Chief Master Sergeant, Air Force Reserve.

Author Disclosure
This work was supported by a research grant from Pfizer and by the Louis Stokes Cleveland VAMC, the VISN 10 Geriatric Research Education and Clinical Center, and the Clinical and Translational Science Collaborative of Cleveland (award UL1TR000439 from the National Center for Advancing Translational Sciences of the National Institutes of Health NIH). The content is the responsibility of the authors and does not represent the official views of the NIH or the VA.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Streptococcus pneumoniae (S pneumoniae), also known as pneumococcus, is a successful human pathogen with significant clinical impact that causes pneumonia and invasive infections, including bacteremia and meningitis.¹ In the preantibiotic era, nearly 80% of patients with bacteremic pneumococcal pneumonia died.² The introduction of sulfas and penicillin in the mid-20th century, subsequent refinements in antibiotic chemotherapy, and improvements in supportive care rendered pneumococcal disease readily treatable, notwithstanding the threat of antibiotic-resistant pneumococcus.³ Despite the availability of effective antibiotic therapy against S pneumoniae, pneumococcal disease remains a significant cause of morbidity and mortality among people with increased susceptibility, such as older adults and those living with chronic illness or immunosuppressive conditions. In developed countries like the U.S., where a growing portion of the population is vulnerable to S pneumoniae by virtue of their advanced age and underlying medical conditions, pneumococcal disease is still an important public health concern.⁴

Penumococcal Vaccines: A Long Time Coming

Vaccination against S pneumoniae has proven an efficacious strategy to reduce the morbidity and mortality associated with this pathogen.⁵ The original efforts to develop a pneumococcal vaccine culminated in 1945 with a vaccine containing pneumococcal capsular polysaccharides, which elicited a protective immune response among U.S. soldiers (Table 1).Subsequent investigations determined that the protective response was specific to pneumococcal disease caused by the 4 pneumococcal capsular serotypes included in the vaccine, that the carrier rate of pneumococcus with the vaccine serotypes decreased by about 50%, and that the incidence of pneumonia from the vaccine serotypes was reduced even in nonimmunized soldiers.⁶ These early observations remain relevant to our contemporary understanding of the impact of pneumococcal vaccination: Protection is limited to serotypes included in the vaccine; the vaccine reduces colonization; and the vaccine leads to herd immunity—the protection of unvaccinated subjects.

Despite this achievement, the use of vaccines as a strategy to combat pneumococcal disease was upstaged in the subsequent decades by the success of antibiotics. Renewed interest in pneumococcal vaccines resulted from the efforts of Robert Austrian, MD, who astutely observed that “highly effective antimicrobial drugs must be supplemented by other measures, both prophylactic and therapeutic, if the significant mortality resulting still from pneumococcal infection is to be reduced.”⁷

After initial studies carried out in South African gold miners, the FDA approved a pneumococcal polysaccharide vaccine (PPSV) against 23 of about 90 circulating serotypes.⁸ The CDC and the Advisory Committee on Immunization Practices (ACIP) initially recommended PPSV-23 for persons perceived to be at high risk for pneumococcal disease, including those with chronic diseases, immunocompromising conditions, and older adults.⁹

Three decades later, the analysis of a large body of available evidence demonstrated the protective effects of PPSV-23 against invasive pneumococcal disease caused by pneumococcal types included in the vaccine, especially bacteremic pneumonia (about 75% reduction, odds ratio [OR] 0.26, 95% confidence interval [CI] 0.14-0.45).¹⁰ A perceived shortcoming of PPSV-23, which some experts dispute, is the lack of definite protection against nonbacteremic pneumococcal pneumonia.¹¹ Studies of nursing home residents demonstrated a 50% reduction in the incidence of both pneumococcal pneumonia and all-cause pneumonia, suggesting that PPSV-23 offers protection against noninvasive pneumococcal pneumonia in specific populations.¹² Additional potential limitations of PPSV-23 include reduced benefit in patients aged > 65 years and waning of protection over time.¹³

Related: Identification and Management of Middle East Respiratory Syndrome

Protein-Conjugate Vaccines

Clearly, the most important limitation of PPSV-23, inherent to all capsular polysaccharide vaccines, is that it does not elicit an immune response in children aged < 2 years. The successful development of a vaccine against Haemophilus influenzae type b (Hib) gave rise to a new generation of pneumococcal vaccines.¹⁴ Specifically, the Hib vaccine covalently binds, or conjugates, the capsule polysaccharide to an antigenic protein, leading to effective T-cell–mediated antibody production in infants and toddlers.

In 2000, children received the first iteration of a protein-conjugate vaccine containing the 7 most relevant pneumococcal serotypes (PCV-7).¹⁵ The effects of PCV-7 on pneumococcal disease have been extraordinary, practically eliminating infections caused by the pneumococcal serotypes included in the vaccine. Immunizing children with PCV-7 also ushered in a fundamental public health benefit for adults aged > 65 years: a reduction of nearly 90% in the incidence of pneumococcal infections caused by serotypes included in the vaccine. By eliciting mucosal immunity, which leads to decreased nasal carriage of the covered pneumococcal strains among children, PVC-7 generates herd immunity, leading to reductions in transmission, colonization, and infection with vaccine serotypes among adults.¹⁶

In 2010, a conjugate vaccine containing 13 serotypes (PCV-13) replaced PCV-7 administration for children.^17,18 The PCV-13 is expected to protect children and the herd from disease caused by 6 additional pneumococcal serotypes, including those that surged as replacement strains, filling the ecologic niche created by PCV-7, such as the epidemiologically relevant serotype 19A.^19,20

Immune-Compromised Adults

One of the shortcomings of PPSV-23 is its lack of efficacy in patients with advanced HIV, a group with an exquisite vulnerability to pneumococcal disease, as demonstrated by a randomized controlled trial of PPSV-23 in African patients with advanced, untreated HIV infection.²¹ Similarly, there is concern that patients with lymphoma, leukemia, multiple myeloma, and Hodgkin disease, also at high risk of pneumococcal infection, do not mount sufficient immune responses to polysaccharide antigens or that these responses are adversely affected by chemotherapy or immune suppressing medications. In these populations, a conjugate vaccine (PCV-7 or PCV-13) may elicit a more robust and durable immune response than a polysaccharide vaccine (PPSV-23). A randomized placebo-controlled trial demonstrated the efficacy of PCV-7 in protecting patients with advanced HIV against pneumococcal disease in Africa.²² Based in part on this observation, the ACIP now recommends the use of PCV-13 in patients with HIV and other immunosuppressing conditions, including chronic renal failure.²³ A direct comparison of the relative protection of PPSV-23 vs PCV-7 in this population has not been performed.

Pneumococcal Vaccines in the Elderly

Although PPSV-23 was widely adopted in the U.S. with the intention to protect adults aged > 65 years from pneumococcal infection, this vaccine did not achieve global appeal. In the Netherlands, for instance, PPSV-23 had very low penetration and was never endorsed by public health authorities. Concerns have been issued about the decreased immunogenicity of a polysaccharide vaccine in older adults, invoking the concept of immune senescence, a term that describes a diminished capacity to mount robust immune responses due to aging.

Protein-conjugated antigens, on the other hand, can elicit a better initial immune response than a polysaccharide vaccine can in an older adult. Whether the effect is sustained and translates to better clinical outcomes remains unknown. The adoption of PCV-13 and the continued role of PPSV-23 in adults aged > 65 years have been examined carefully by experts.^24,25

Dutch investigators organized the Community Acquired Pneumonia Immunization Trial in Adults (CAPITA), randomizing about 85,000 adults aged > 65 years toreceive eitherPCV-13 or a placebo.²⁶ After 3 years of observation, the occurrence of invasive pneumococcal disease (caused by the vaccine serotypes) decreased by 85% in participants who received the vaccine compared with those who received a placebo. Additionally, the incidence of pneumococcal pneumonia (caused by the vaccine serotypes) decreased by 45% (Table 2). The findings of this trial largely inform the recommendation issued by APIC in October 2014 to administer PCV-13 vaccine to adults aged > 65 years.²⁷

Polysaccharide Vaccine in Adults

A crucial limitation of the CAPITA study is that it does not provide a head-to-head comparison with PPSV-23. Similarly, the recommendation to use PCV-13 in patients with immune system compromising conditions did not arise from a direct comparison between PCV-13 and PPSV-23. In both groups of patients, ACIP continues to recommend PPSV-23 to extend protection to the 10 additional pneumococcal serotypes not included in PCV-13.²⁸ Additionally, ACIP maintains its long-standing recommendation to administer PPSV-23 to adults aged < 65 years who have diabetes mellitus, or chronic liver, lung, or heart disease, or those who are tobacco smokers or misuse alcohol.

Notably, the order in which the vaccines are administered may influence their effectiveness. In adults aged 50 to 64 years, initial vaccination with PCV-13 followed by PPSV-23 lead to a better antipneumococcal immune response.^14,29 Specifically, PCV-13 enhanced the response to subsequent administration of PPSV-23, whereas an initial PPSV-23 vaccine resulted in a decreased response to subsequent administration of PCV-13. After a 1-year interval between vaccines, immune titers to a second vaccination were not inferior.³⁰ All the above considerations have shaped the ACIP recommendations for the administration of pneumococcal vaccines.^27,31 Pneumococcal vaccination was a straightforward exercise when PPSV-23 was the only vaccine available for adults. The advent of PCV-13, however, renders pneumococcal immunization of adults more complicated (Figure 1).

Improving Vaccination Rates

The complex recommendations to vaccinate adults against pneumococcus may reveal obstacles to effective pneumococcal vaccination programs.³² From the perspective of health care institutions, the adoption of the new pneumococcal vaccine implies an added cost. A pneumococcal vaccination strategy that incorporates PCV-13 may be cost-effective, at least under certain parameters.^33-35 There remains, however, the issue of affordability and the opportunity cost—the need to decrease funding for other health care programs to accommodate an increased budget for pneumococcal vaccination. Logistically, maintaining a reliable supply of vaccines to meet the demand of practitioners at various sites requires careful planning. Consequently, tracking vaccinated and unvaccinated patients and carefully coordinating among clinical providers, nurses, and pharmacists are essential.

From the perspective of clinical providers, an additional pneumococcal vaccine complicating the vaccination schedule for adults represents an increased burden. Providers will need information to reach their own conclusions regarding the rationale behind the development and use of pneumococcal conjugate vaccines, the existing and evolving recommendations from public health authorities, and the strengths and limitations of the evidence supporting the use of pneumococcal vaccines. Otherwise, providers may find it difficult to incorporate new data and guidelines supporting pneumococcal vaccination into their decision-making (Boxes 1-4). An additional and formidable challenge is to carve out time during an already busy clinical encounter to discuss pneumococcal vaccines and other immunizations.³⁶

Related: The Importance of an Antimicrobial Stewardship Program

Similarly, older adults or patients with chronic health conditions may not recognize the important role that vaccines can play in their health maintenance and are likely to prioritize other issues during medical visits. It is not obvious for patients that multiple vaccines may be necessary to prevent pneumococcal disease. Furthermore, many patients, not unreasonably, may assume that their yearly influenza vaccine is sufficiently protective against pneumonia. Therefore, patients need to be educated about the rationale behind pneumococcal conjugate vaccination. Ultimately, access to immunization—the opportunity for patients to have an encounter with their providers and with the health care system that results in the administration of an appropriate vaccine—will determine whether goals for pneumococcal vaccination are achieved.

The evolving landscape for the implementation of pneumococcal vaccines creates the need to develop, implement, and refine organizational changes to adhere to the new guidelines for the use of PCV-13 and PPSV-23 vaccines. These interventions, if effective, may help improve pneumococcal vaccination coverage among adults (Table 3).

Harnessing the Power of the EMR

Interventions to improve the adherence to pneumococcal vaccination guidelines begin by identifying persons eligible for vaccination based on their age, their vaccination status (ie, persons previously unvaccinated or due for vaccination according to the recommended schedule), or the presence of medical conditions conferring high risk for pneumococcal disease. This, in turn, depends on adequate documentation of patients’ underlying medical diagnosis, as well as up-to-date records of vaccine administration to patients.

Health care systems possessing a mature and sophisticated electronic medical record (EMR), such as the VHA Computerized Patient Record System (CPRS), are in a good position to wield such information to plan, implement, and assure the quality of activities designed to improve pneumococcal vaccination rates. An analysis of the proportion of veterans in VISN 10 who received pneumococcal vaccination revealed that even with the advantages of a robust EMR and a highly developed infrastructure devoted to primary care, pneumococcal vaccine coverage remains below the 60% target goal, although well above national averages (Figure 2).^37,38

Standing Orders

Standing orders make it possible for nurses and pharmacists to administer vaccines according to a preestablished protocol without a physician’s direct evaluation of each patient. Standing orders are a versatile intervention, with a record of effective implementation in both outpatient and inpatient facilities, in acute-care and long-term care facilities, and in most instances where patients interact with the health care system. Based on strong scientific evidence, ACIP recommends the adoption of standing-order programs to improve pneumococcal vaccination rates among adults.³⁹ Indeed, standing-order programs may prove a very effective intervention to fulfill the recommendation to administer PCV-13 to adults aged > 65 years.

The Virtual Vaccination Clinic

Unfortunately, with 2 vaccines that have to be administered at different times to various groups of patients at risk, the current state of pneumococcal vaccination may be too complex to be readily reduced to a comprehensible set of standing orders. An innovative way to realize the benefit of standing orders is to target high-risk groups for pneumococcal disease who are eligible for vaccination by selecting them using the EMR and entering standing orders tailored to their specific vaccination needs. The selection of patients according to comorbidities and vaccination status and the determination of the appropriate pneumococcal vaccine takes place in the context of a “virtual” vaccination clinic.⁴⁰

Enhancing Vaccinations


Further improvement in pneumococcal vaccination rates are likely to result from interventions that increase the demand for vaccines among patients and practitioners. Efforts to disseminate information and provide advice regarding pneumococcal vaccination are likely to result in patients seeking and clinicians offering the appropriate vaccine. Similarly, interventions to enhance the supply of vaccines at the point of care may reduce barriers that patients might encounter when attempting to receive vaccinations.

Another set of system-based interventions that can assist clinicians in making timely and appropriate vaccination decisions are EMR reminders, especially those targeting patients at high risk for pneumococcal disease because of underlying illnesses.⁴¹ Previous experience with pneumococcal vaccination in patients aged > 65 years indicates that EMR reminders facilitate improvements in vaccination. The involvement of a panel manager who coordinated with the primary care provider and contacted patients directly augmented the effect of the EMR reminder by 25%.⁴²

Related: Venous Thromboembolism Prophylaxis in Acutely Ill Veterans With Respiratory Disease

Patient reminder and recall systems also demonstrated effectiveness in improving immunization rates.⁴³ In certain groups, notification of patients has been achieved through electronic methods, such as short text messaging or e-mail.⁴⁴ Determining which interventions within a bundle are essential may be impossible, because the various interventions reinforce one another, and the likelihood of patients benefitting from at least one of the activities increases when multiple interventions are administered together. Therefore, the Task Force on Community Preventive Services supports combining provider reminder systems with education and other measures that encourage use of vaccines in patients and providers.⁴⁵

Box answer key: 1: A; 2: A; 3: B; 4: A.

Conclusion

The increasing role of vaccines in the health maintenance of adults represents a change in paradigm for primary care and specialty providers. Physicians must assess the value and limitations of vaccines and find time to discuss immunizations with their adult patients. Health care systems can increase opportunities for vaccination and facilitate encounters that result in vaccination by educating patients and health care personnel and through the innovative use of reminders and standing orders in the EMR. Undertaking these activities may limit the burden of pneumococcal disease, an important cause of morbidity and mortality in adults that is preventable through immunization.

Acknowledgements
This work is dedicated to the memory of John M. Rieger, Information Technology Specialist at the Cleveland VAMC and Chief Master Sergeant, Air Force Reserve.

Author Disclosure
This work was supported by a research grant from Pfizer and by the Louis Stokes Cleveland VAMC, the VISN 10 Geriatric Research Education and Clinical Center, and the Clinical and Translational Science Collaborative of Cleveland (award UL1TR000439 from the National Center for Advancing Translational Sciences of the National Institutes of Health NIH). The content is the responsibility of the authors and does not represent the official views of the NIH or the VA.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Streptococcus pneumoniae (S pneumoniae), also known as pneumococcus, is a successful human pathogen with significant clinical impact that causes pneumonia and invasive infections, including bacteremia and meningitis.¹ In the preantibiotic era, nearly 80% of patients with bacteremic pneumococcal pneumonia died.² The introduction of sulfas and penicillin in the mid-20th century, subsequent refinements in antibiotic chemotherapy, and improvements in supportive care rendered pneumococcal disease readily treatable, notwithstanding the threat of antibiotic-resistant pneumococcus.³ Despite the availability of effective antibiotic therapy against S pneumoniae, pneumococcal disease remains a significant cause of morbidity and mortality among people with increased susceptibility, such as older adults and those living with chronic illness or immunosuppressive conditions. In developed countries like the U.S., where a growing portion of the population is vulnerable to S pneumoniae by virtue of their advanced age and underlying medical conditions, pneumococcal disease is still an important public health concern.⁴

Penumococcal Vaccines: A Long Time Coming

Vaccination against S pneumoniae has proven an efficacious strategy to reduce the morbidity and mortality associated with this pathogen.⁵ The original efforts to develop a pneumococcal vaccine culminated in 1945 with a vaccine containing pneumococcal capsular polysaccharides, which elicited a protective immune response among U.S. soldiers (Table 1).Subsequent investigations determined that the protective response was specific to pneumococcal disease caused by the 4 pneumococcal capsular serotypes included in the vaccine, that the carrier rate of pneumococcus with the vaccine serotypes decreased by about 50%, and that the incidence of pneumonia from the vaccine serotypes was reduced even in nonimmunized soldiers.⁶ These early observations remain relevant to our contemporary understanding of the impact of pneumococcal vaccination: Protection is limited to serotypes included in the vaccine; the vaccine reduces colonization; and the vaccine leads to herd immunity—the protection of unvaccinated subjects.

Despite this achievement, the use of vaccines as a strategy to combat pneumococcal disease was upstaged in the subsequent decades by the success of antibiotics. Renewed interest in pneumococcal vaccines resulted from the efforts of Robert Austrian, MD, who astutely observed that “highly effective antimicrobial drugs must be supplemented by other measures, both prophylactic and therapeutic, if the significant mortality resulting still from pneumococcal infection is to be reduced.”⁷

After initial studies carried out in South African gold miners, the FDA approved a pneumococcal polysaccharide vaccine (PPSV) against 23 of about 90 circulating serotypes.⁸ The CDC and the Advisory Committee on Immunization Practices (ACIP) initially recommended PPSV-23 for persons perceived to be at high risk for pneumococcal disease, including those with chronic diseases, immunocompromising conditions, and older adults.⁹

Three decades later, the analysis of a large body of available evidence demonstrated the protective effects of PPSV-23 against invasive pneumococcal disease caused by pneumococcal types included in the vaccine, especially bacteremic pneumonia (about 75% reduction, odds ratio [OR] 0.26, 95% confidence interval [CI] 0.14-0.45).¹⁰ A perceived shortcoming of PPSV-23, which some experts dispute, is the lack of definite protection against nonbacteremic pneumococcal pneumonia.¹¹ Studies of nursing home residents demonstrated a 50% reduction in the incidence of both pneumococcal pneumonia and all-cause pneumonia, suggesting that PPSV-23 offers protection against noninvasive pneumococcal pneumonia in specific populations.¹² Additional potential limitations of PPSV-23 include reduced benefit in patients aged > 65 years and waning of protection over time.¹³

Related: Identification and Management of Middle East Respiratory Syndrome

Protein-Conjugate Vaccines

Clearly, the most important limitation of PPSV-23, inherent to all capsular polysaccharide vaccines, is that it does not elicit an immune response in children aged < 2 years. The successful development of a vaccine against Haemophilus influenzae type b (Hib) gave rise to a new generation of pneumococcal vaccines.¹⁴ Specifically, the Hib vaccine covalently binds, or conjugates, the capsule polysaccharide to an antigenic protein, leading to effective T-cell–mediated antibody production in infants and toddlers.

In 2000, children received the first iteration of a protein-conjugate vaccine containing the 7 most relevant pneumococcal serotypes (PCV-7).¹⁵ The effects of PCV-7 on pneumococcal disease have been extraordinary, practically eliminating infections caused by the pneumococcal serotypes included in the vaccine. Immunizing children with PCV-7 also ushered in a fundamental public health benefit for adults aged > 65 years: a reduction of nearly 90% in the incidence of pneumococcal infections caused by serotypes included in the vaccine. By eliciting mucosal immunity, which leads to decreased nasal carriage of the covered pneumococcal strains among children, PVC-7 generates herd immunity, leading to reductions in transmission, colonization, and infection with vaccine serotypes among adults.¹⁶

In 2010, a conjugate vaccine containing 13 serotypes (PCV-13) replaced PCV-7 administration for children.^17,18 The PCV-13 is expected to protect children and the herd from disease caused by 6 additional pneumococcal serotypes, including those that surged as replacement strains, filling the ecologic niche created by PCV-7, such as the epidemiologically relevant serotype 19A.^19,20

Immune-Compromised Adults

One of the shortcomings of PPSV-23 is its lack of efficacy in patients with advanced HIV, a group with an exquisite vulnerability to pneumococcal disease, as demonstrated by a randomized controlled trial of PPSV-23 in African patients with advanced, untreated HIV infection.²¹ Similarly, there is concern that patients with lymphoma, leukemia, multiple myeloma, and Hodgkin disease, also at high risk of pneumococcal infection, do not mount sufficient immune responses to polysaccharide antigens or that these responses are adversely affected by chemotherapy or immune suppressing medications. In these populations, a conjugate vaccine (PCV-7 or PCV-13) may elicit a more robust and durable immune response than a polysaccharide vaccine (PPSV-23). A randomized placebo-controlled trial demonstrated the efficacy of PCV-7 in protecting patients with advanced HIV against pneumococcal disease in Africa.²² Based in part on this observation, the ACIP now recommends the use of PCV-13 in patients with HIV and other immunosuppressing conditions, including chronic renal failure.²³ A direct comparison of the relative protection of PPSV-23 vs PCV-7 in this population has not been performed.

Pneumococcal Vaccines in the Elderly

Although PPSV-23 was widely adopted in the U.S. with the intention to protect adults aged > 65 years from pneumococcal infection, this vaccine did not achieve global appeal. In the Netherlands, for instance, PPSV-23 had very low penetration and was never endorsed by public health authorities. Concerns have been issued about the decreased immunogenicity of a polysaccharide vaccine in older adults, invoking the concept of immune senescence, a term that describes a diminished capacity to mount robust immune responses due to aging.

Protein-conjugated antigens, on the other hand, can elicit a better initial immune response than a polysaccharide vaccine can in an older adult. Whether the effect is sustained and translates to better clinical outcomes remains unknown. The adoption of PCV-13 and the continued role of PPSV-23 in adults aged > 65 years have been examined carefully by experts.^24,25

Dutch investigators organized the Community Acquired Pneumonia Immunization Trial in Adults (CAPITA), randomizing about 85,000 adults aged > 65 years toreceive eitherPCV-13 or a placebo.²⁶ After 3 years of observation, the occurrence of invasive pneumococcal disease (caused by the vaccine serotypes) decreased by 85% in participants who received the vaccine compared with those who received a placebo. Additionally, the incidence of pneumococcal pneumonia (caused by the vaccine serotypes) decreased by 45% (Table 2). The findings of this trial largely inform the recommendation issued by APIC in October 2014 to administer PCV-13 vaccine to adults aged > 65 years.²⁷

Polysaccharide Vaccine in Adults

A crucial limitation of the CAPITA study is that it does not provide a head-to-head comparison with PPSV-23. Similarly, the recommendation to use PCV-13 in patients with immune system compromising conditions did not arise from a direct comparison between PCV-13 and PPSV-23. In both groups of patients, ACIP continues to recommend PPSV-23 to extend protection to the 10 additional pneumococcal serotypes not included in PCV-13.²⁸ Additionally, ACIP maintains its long-standing recommendation to administer PPSV-23 to adults aged < 65 years who have diabetes mellitus, or chronic liver, lung, or heart disease, or those who are tobacco smokers or misuse alcohol.

Notably, the order in which the vaccines are administered may influence their effectiveness. In adults aged 50 to 64 years, initial vaccination with PCV-13 followed by PPSV-23 lead to a better antipneumococcal immune response.^14,29 Specifically, PCV-13 enhanced the response to subsequent administration of PPSV-23, whereas an initial PPSV-23 vaccine resulted in a decreased response to subsequent administration of PCV-13. After a 1-year interval between vaccines, immune titers to a second vaccination were not inferior.³⁰ All the above considerations have shaped the ACIP recommendations for the administration of pneumococcal vaccines.^27,31 Pneumococcal vaccination was a straightforward exercise when PPSV-23 was the only vaccine available for adults. The advent of PCV-13, however, renders pneumococcal immunization of adults more complicated (Figure 1).

Improving Vaccination Rates

The complex recommendations to vaccinate adults against pneumococcus may reveal obstacles to effective pneumococcal vaccination programs.³² From the perspective of health care institutions, the adoption of the new pneumococcal vaccine implies an added cost. A pneumococcal vaccination strategy that incorporates PCV-13 may be cost-effective, at least under certain parameters.^33-35 There remains, however, the issue of affordability and the opportunity cost—the need to decrease funding for other health care programs to accommodate an increased budget for pneumococcal vaccination. Logistically, maintaining a reliable supply of vaccines to meet the demand of practitioners at various sites requires careful planning. Consequently, tracking vaccinated and unvaccinated patients and carefully coordinating among clinical providers, nurses, and pharmacists are essential.

From the perspective of clinical providers, an additional pneumococcal vaccine complicating the vaccination schedule for adults represents an increased burden. Providers will need information to reach their own conclusions regarding the rationale behind the development and use of pneumococcal conjugate vaccines, the existing and evolving recommendations from public health authorities, and the strengths and limitations of the evidence supporting the use of pneumococcal vaccines. Otherwise, providers may find it difficult to incorporate new data and guidelines supporting pneumococcal vaccination into their decision-making (Boxes 1-4). An additional and formidable challenge is to carve out time during an already busy clinical encounter to discuss pneumococcal vaccines and other immunizations.³⁶

Related: The Importance of an Antimicrobial Stewardship Program

Similarly, older adults or patients with chronic health conditions may not recognize the important role that vaccines can play in their health maintenance and are likely to prioritize other issues during medical visits. It is not obvious for patients that multiple vaccines may be necessary to prevent pneumococcal disease. Furthermore, many patients, not unreasonably, may assume that their yearly influenza vaccine is sufficiently protective against pneumonia. Therefore, patients need to be educated about the rationale behind pneumococcal conjugate vaccination. Ultimately, access to immunization—the opportunity for patients to have an encounter with their providers and with the health care system that results in the administration of an appropriate vaccine—will determine whether goals for pneumococcal vaccination are achieved.

The evolving landscape for the implementation of pneumococcal vaccines creates the need to develop, implement, and refine organizational changes to adhere to the new guidelines for the use of PCV-13 and PPSV-23 vaccines. These interventions, if effective, may help improve pneumococcal vaccination coverage among adults (Table 3).

Harnessing the Power of the EMR

Interventions to improve the adherence to pneumococcal vaccination guidelines begin by identifying persons eligible for vaccination based on their age, their vaccination status (ie, persons previously unvaccinated or due for vaccination according to the recommended schedule), or the presence of medical conditions conferring high risk for pneumococcal disease. This, in turn, depends on adequate documentation of patients’ underlying medical diagnosis, as well as up-to-date records of vaccine administration to patients.

Health care systems possessing a mature and sophisticated electronic medical record (EMR), such as the VHA Computerized Patient Record System (CPRS), are in a good position to wield such information to plan, implement, and assure the quality of activities designed to improve pneumococcal vaccination rates. An analysis of the proportion of veterans in VISN 10 who received pneumococcal vaccination revealed that even with the advantages of a robust EMR and a highly developed infrastructure devoted to primary care, pneumococcal vaccine coverage remains below the 60% target goal, although well above national averages (Figure 2).^37,38

Standing Orders

Standing orders make it possible for nurses and pharmacists to administer vaccines according to a preestablished protocol without a physician’s direct evaluation of each patient. Standing orders are a versatile intervention, with a record of effective implementation in both outpatient and inpatient facilities, in acute-care and long-term care facilities, and in most instances where patients interact with the health care system. Based on strong scientific evidence, ACIP recommends the adoption of standing-order programs to improve pneumococcal vaccination rates among adults.³⁹ Indeed, standing-order programs may prove a very effective intervention to fulfill the recommendation to administer PCV-13 to adults aged > 65 years.

The Virtual Vaccination Clinic

Unfortunately, with 2 vaccines that have to be administered at different times to various groups of patients at risk, the current state of pneumococcal vaccination may be too complex to be readily reduced to a comprehensible set of standing orders. An innovative way to realize the benefit of standing orders is to target high-risk groups for pneumococcal disease who are eligible for vaccination by selecting them using the EMR and entering standing orders tailored to their specific vaccination needs. The selection of patients according to comorbidities and vaccination status and the determination of the appropriate pneumococcal vaccine takes place in the context of a “virtual” vaccination clinic.⁴⁰

Enhancing Vaccinations


Further improvement in pneumococcal vaccination rates are likely to result from interventions that increase the demand for vaccines among patients and practitioners. Efforts to disseminate information and provide advice regarding pneumococcal vaccination are likely to result in patients seeking and clinicians offering the appropriate vaccine. Similarly, interventions to enhance the supply of vaccines at the point of care may reduce barriers that patients might encounter when attempting to receive vaccinations.

Another set of system-based interventions that can assist clinicians in making timely and appropriate vaccination decisions are EMR reminders, especially those targeting patients at high risk for pneumococcal disease because of underlying illnesses.⁴¹ Previous experience with pneumococcal vaccination in patients aged > 65 years indicates that EMR reminders facilitate improvements in vaccination. The involvement of a panel manager who coordinated with the primary care provider and contacted patients directly augmented the effect of the EMR reminder by 25%.⁴²

Related: Venous Thromboembolism Prophylaxis in Acutely Ill Veterans With Respiratory Disease

Patient reminder and recall systems also demonstrated effectiveness in improving immunization rates.⁴³ In certain groups, notification of patients has been achieved through electronic methods, such as short text messaging or e-mail.⁴⁴ Determining which interventions within a bundle are essential may be impossible, because the various interventions reinforce one another, and the likelihood of patients benefitting from at least one of the activities increases when multiple interventions are administered together. Therefore, the Task Force on Community Preventive Services supports combining provider reminder systems with education and other measures that encourage use of vaccines in patients and providers.⁴⁵

Box answer key: 1: A; 2: A; 3: B; 4: A.

Conclusion

The increasing role of vaccines in the health maintenance of adults represents a change in paradigm for primary care and specialty providers. Physicians must assess the value and limitations of vaccines and find time to discuss immunizations with their adult patients. Health care systems can increase opportunities for vaccination and facilitate encounters that result in vaccination by educating patients and health care personnel and through the innovative use of reminders and standing orders in the EMR. Undertaking these activities may limit the burden of pneumococcal disease, an important cause of morbidity and mortality in adults that is preventable through immunization.

Acknowledgements
This work is dedicated to the memory of John M. Rieger, Information Technology Specialist at the Cleveland VAMC and Chief Master Sergeant, Air Force Reserve.

Author Disclosure
This work was supported by a research grant from Pfizer and by the Louis Stokes Cleveland VAMC, the VISN 10 Geriatric Research Education and Clinical Center, and the Clinical and Translational Science Collaborative of Cleveland (award UL1TR000439 from the National Center for Advancing Translational Sciences of the National Institutes of Health NIH). The content is the responsibility of the authors and does not represent the official views of the NIH or the VA.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Patients with known cardiovascular (CV) disease are at greater risk for CV events.¹ Hydroxymethylglutaryl-CoA (HMG Co-A) reductase inhibitors, or statins, have been shown to reduce CV events and to reduce all-cause mortality.^1,2 Thus, these agents should be a standard approach to secondary prevention of CV events.^1,2 Although the main function of statins is to lower total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) levels, trials have shown that other lipid-lowering agents have reduced the incidence of CV events but have failed to show any difference in mortality.¹ Thus, the therapeutic effects of statins may be a result of “pleiotropic” effects in addition to a reduction in LDL-C.³

As a result, prescribing practices and professional society guidelines have deferred to statins as a first-line choice for lipid-lowering therapy.¹ Although each statin varies in its ability to lower LDL-C and inhibit HMG Co-A reductase, as a class, statins have been proven to be safe and efficacious in reducing LDL-C, decreasing risk of coronary artery disease, and decreasing mortality.

The 2013 American College of Cardiology (ACC) and American Heart Association (AHA) guideline on the treatment of blood cholesterol to reduce CV risk in adults resulted in a major shift in clinical practice recommendations. The focus of treatment has changed from LDL-C and TC goals to stratifying patients to either high-intensity or moderate-intensity statin therapy, based on their comorbidities and risk of atherosclerotic CV disease (ASCVD).¹ Primary prevention of CV disease has been proposed for patients with diabetes mellitus aged 40 to 75 years, familial hypercholesterolemia (LDL-C > 190 mg/dL), and for patients with an ASCVD risk score > 7.5%. Secondary prevention has been proposed for all patients with a history of ASCVD. Among the available choices for intensive statin therapy, the 2 most potent regimens are atorvastatin (40-80 mg) and rosuvastatin (20-40 mg) daily. The ACC/AHA guideline recommends high-potency therapy with either rosuvastatin or atorvastatin with equal preference.¹

Related: New Incentives for Helping Prevent Heart Disease

Statin therapy is generally well tolerated; however, the use of statins is not without risk of adverse drug reactions (ADRs). Skeletal muscle discomfort has been reported in 4% to 10% of patients taking either atorvastatin or rosuvastatin.^4,5 Liver enzyme abnormalities are less common, having been reported in only about 2% to 3% of patients taking either atorvastatin or rosuvastatin.^4,5 Although muscle-related intolerance and liver enzyme abnormalities are considered to be class effects, research speculates that the therapeutic and safety effects of statins may differ, based on tissue solubility.^2,6 Variability in the myotoxic and hepatotoxic effects of statins has been attributed to differences in tissue solubility, hypothesizing that lipophilic statins are more easily taken up into myocytes and hepatocytes, resulting in an increase in toxic effects.^2,6

This study assessed differences in therapeutic and safety endpoints resulting from the recent interchange from rosuvastatin to atorvastatin within the North Florida/South Georgia Veterans Health System (NF/SGVHS). Although both these agents are high-potency HMG Co-A reductase inhibitors and share a mechanism of action, they have pharmacokinetic differences, including a key difference in tissue solubility (Table 1).

With the availability of low-cost generic atorvastatin in early 2012, the VA Medical Advisory Panel and Pharmacy Benefits Management (VA MAP/PBM) added atorvastatin to the VA National Formulary as the preferred high-potency statin.^7,8 Before October 2012, rosuvastatin had been the preferred high-potency statin within the VA. With support from VA MAP/PBM leadership, NF/SGVHS instituted an interchange from rosuvastatin to atorvastatin for cost-savings purposes.⁹

Before the interchange, the records of patients were reviewed to determine whether justification existed for continued use of rosuvastatin. Patients were converted to atorvastatin if deemed appropriate and received education and consultation through direct patient contact or a letter regarding the interchange. Justifications for continued use of rosuvastatin included documentation of an atorvastatin ADR, active liver disease, or patients taking cyclosporine or certain protease inhibitors.⁸

The objective of this retrospective evaluation was to assess the efficacy and safety of the interchange from rosuvastatin to atorvastatin within NF/SGVHS. The results of this review are helpful to confirm the efficacy and safety of the interchange and identify any differences in efficacy and safety that may have occurred as a result of the interchange to a different high-potency statin. For this review, statin efficacy was assessed via review of pre- and postinterchange lipid panel values, assessing for a significant difference between equipotent atorvastatin and rosuvastatin therapy. Similarly, safety was assessed by analysis of pre- and postinterchange liver enzyme panels and assessing for significant differences as a result of the interchange.

Methods

The therapeutic interchange was conducted within the NF/SGVHS, which provides patient care at hospitals in Gainesville, Florida, and Lake City, Florida, and 11 outpatient clinics located throughout North Florida and South Georgia. Like other VA facilities, NF/SGVHS uses Computerized Patient Records System (CPRS) to electronically integrate all clinical patient information, including medical progress notes, consults, admission and discharge summaries, allergies and ADRs, patient problem lists (diagnoses), vital signs, medication orders, and laboratory test results. Approval to conduct this study was granted by the University of Florida Investigational Review Board and the Research and Development Committee at NF/SGVHS.

Therapeutic Interchange Process

Interchange from rosuvastatin to atorvastatin was expected to provide about $643,000 annually in drug cost savings to NF/SGVHS while providing equivalent therapy. The cost for a 30-day supply of rosuvastatin was about $22.56 at the time of interchange, and a 30-day supply of generic atorvastatin was $1.77.The interchange from rosuvastatin to atorvastatin was approved by the Pharmacy and Therapeutics Committee Meeting on August 8, 2012, and began shortly thereafter. Interchanges were halted temporarily in November 2012 due to a shortage of manufacturer supply, but the process fully resumed in January 2013 once the drug shortage resolved. Direction was provided to VA facilities by a guidance letter issued through PBM leadership.⁸ The interchange used a standard interchange guide to complete the process (Table 2).

Posttherapeutic Interchange Analysis

Researchers conducted an internal pharmacy computerizedprescription records search to identify VA outpatients who were converted from rosuvastatin to atorvastatin from February 1, 2012, to August 1, 2013. A total of 202 patients were randomly selected and included in this retrospective chart review. Investigators analyzed data points for safety and efficacy, including liver function tests (LFTs), lipid panels, and ADR reports. This information was obtained from laboratory data, vital signs, allergy information, ADR data, and progress notes using CPRS. Two sets of laboratory data were obtained for research purposes, the most recent laboratory values pre-interchange and the most recent laboratory values postinterchange to atorvastatin.

Investigators determined whether patients were converted to an equivalent dose of atorvastatin through an assessment of the most recent dosage of rosuvastatin before the interchange and the dosage of atorvastatin postinterchange. Researchers also analyzed refill history and interacting medications to assess possible confounding factors. Medication adherence was assessed via refill history. Medication adherence was defined as a medication possession ratio of at least 70%, which correlated to receipt of 3 or more 90-day supplies in the year prior to interchange. The above data were collected via retrospective chart review and entered into a spreadsheet.

Statistics

All identifying information was removed from the data set prior to statistical analysis. The data analysis for this project was generated using SAS/STAT software, version 9.3 of the SAS System for Linux x64 (SAS Institute, Cary, North Carolina).

Researchers assumed an equal variance in preinterchange and postinterchange lipid and liver panel values. Preinterchange and postinterchange lipid values (LDL-C, HDL-C, total cholesterol [TC], and triglycerides [TGs]) and liver values (aspartate aminotransferase [AST], alanine aminotransferase [ALT], alkaline phosphatase [ALP], and creatinine phosphokinase [CPK]) were analyzed by paired t test. All values were reported as mean (SD), and significance was defined as P < .05.

Results

More than 6,000 veterans within the NF/SGVHS were identified as eligible for the interchange. Of those who were converted, 202 patient records were randomly selected and reviewed. Patient population characteristics are summarized in Table 3. Most patients were aged > 65 years (61.4%) with an average body mass index (BMI) of 32.4. Most patients were converted to the correct corresponding dose of atorvastatin (82.7%) and achieved adherence with statin therapy (84.2%). There was no difference in pre- and postinterchange adherence detected as a result of this review.

Related: New Guideline on Dyslipidemia: Less Is More

Adverse drug reactions were documented in 16 cases, accounting for 8% of the study population. The most commonly reported ADR was myalgia or arthralgia, which was found in 10 cases (5%). Other ADRs identified in this retrospective review included treatment failure, nausea, vomiting, abdominal discomfort, abnormal liver enzymes, nasopharyngitis, and pruritis (Table 4). Of note, treatment failure was determined on a case-by-case basis but was generally defined as a failure to reach LDL-C goal (< 100 mg/dL or < 70 mg/dL), despite titration of atorvastatin. Interacting medications were identified in 30.7% of patients; however, no reported ADRs were associated with interacting medications. The most common drug interaction was concomitant niacin, followed by antiarrhythmics (ie, amiodarone, diltiazem, etc), and fibrates (ie, gemfibrozil, fenofibrate). All potentially interacting medications identified in this retrospective chart review are compiled in Table 5.

No significant difference between mean pre- and postinterchange lipid panel values was identified in this retrospective chart review (Table 6). In addition, no significant difference was detected in pre- and post-interchange AST, ALT, and CPK values (Table 7). However, a statistically significant increase in ALP was detected, with a mean ALP of 73.33 IU/L prior to interchange and 83.64 IU/L postinterchange (P < .0001).

Discussion

The goal of this retrospective observation was to ensure that safety and efficacy were not compromised as a result of this cost-saving therapeutic interchange. No differences in liver enzymes (safety) and lipid control (effectiveness) were observed in this study. There were no statistically significant changes to the lipid panel or liver panel detected with the exception of ALP. The reason for this statistically significant increase is unknown; however, it may support the hypothesis of variation in hepatocellular effects between the statins due to lipophilic properties.^3,6 In general, liver enzymes can be affected by extrahepatic functions. Serum ALP and other liver enzymes can be affected by bone disease, abdominal adiposity, alcoholism, and other concomitant diseases.¹⁰ No comorbid conditions were assessed, thus differences in liver enzymes may not be fully attributable to statin therapy. This retrospective review found no clinically significant effect correlated with the increase in ALP.

The results of this analysis are congruent with similar therapeutic interchange studies, which resulted in cost savings without compromising safety or efficacy.^9,11 Unlike other therapeutic interchange studies, this study analyzed both safety and lipid-lowering efficacy outcomes, instead of focusing solely on changes in LDL-C lowering, total cost savings, and/or adherence.^12,13 By including the entire lipid panel and liver panel into the review, this study conducted a more inclusive review of interchangeability with statins, addressing issues such as HDL-C lowering, TG changes, and liver enzyme fluctuation on conversion. There had not been a sufficient time to assess efficacy in terms of CV outcomes.

Related: Poor Outcomes for African Americans in Cardiac Rehabilitation

Two adverse events alluded to therapeutic failure as a reason for discontinuing atorvastatin. In the previous ATP III lipid guidelines, therapeutic failure was achieved when patients did not reach their LDL-C goals despite appropriate titration of statin therapy.² However, the ACC/AHA lipid guidelines have done away with lipid goals as a measurement of treatment therapy, focusing rather on evidence-based high- or moderate-intensity statin therapy that has been proven in clinical trials to reduce mortality and CV events.¹ Although measurement of efficacy via lipid panel values is no longer a guideline recommendation, the results of this chart review have shown no difference in lipid values as a result of the interchange, confirming the interchangeability of rosuvastatin and atorvastatin at their equivalent doses.

Limitations

The interchange of rosuvastatin to atorvastatin was a policy change affecting all patients within the NF/SGVHS. In order to reflect true population data and more accurately predict the effects of such a policy change, this study used intention-to-treat analysis, including all patients, even patients who were found to be nonadherent. This study is also limited by sample size (N = 202). Additionally, the generalizability of these findings may be limited. The study population was mostly males aged > 65 years with an average BMI of 32.4. Researchers did not compile comorbidity, race, or concomitant medication data. Additionally, the duration of statin therapy prior to laboratory value collection was undefined.

A retrospective chart review lends itself to limitations in data collection. Medication adherence is a factor that is assumed to have a significant effect on the results of this interchange. In this review, adherence was assessed via refill history. Researchers were unable to confirm actual consumption of the medication.

Additionally, researchers did not analyze comorbid conditions, which may have had an effect on lipid panel and liver panel values. For those veterans who discontinued atorvastatin therapy, the reason for discontinuation was often not documented. Thus, researchers were unable to assess reasons for discontinuation.

Conclusion

The results generated from a review of the therapeutic exchange of rosuvastatin to atorvastatin within a veteran population affirm that the interchange was not associated with any differences in safety or lipid control, but did result in significant drug cost savings. This study provides support for health care systems considering therapeutic interchange with high-intensity statins safely and effectively.

Acknowledgements
This material is the result of work supported with resources and the use of facilities at the North Florida/South Georgia Veterans Health System in Gainesville, Florida. The authors would like to acknowledge Kim Hoang, PharmD, for her contributions to this project.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer


The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Patients with known cardiovascular (CV) disease are at greater risk for CV events.¹ Hydroxymethylglutaryl-CoA (HMG Co-A) reductase inhibitors, or statins, have been shown to reduce CV events and to reduce all-cause mortality.^1,2 Thus, these agents should be a standard approach to secondary prevention of CV events.^1,2 Although the main function of statins is to lower total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) levels, trials have shown that other lipid-lowering agents have reduced the incidence of CV events but have failed to show any difference in mortality.¹ Thus, the therapeutic effects of statins may be a result of “pleiotropic” effects in addition to a reduction in LDL-C.³

As a result, prescribing practices and professional society guidelines have deferred to statins as a first-line choice for lipid-lowering therapy.¹ Although each statin varies in its ability to lower LDL-C and inhibit HMG Co-A reductase, as a class, statins have been proven to be safe and efficacious in reducing LDL-C, decreasing risk of coronary artery disease, and decreasing mortality.

The 2013 American College of Cardiology (ACC) and American Heart Association (AHA) guideline on the treatment of blood cholesterol to reduce CV risk in adults resulted in a major shift in clinical practice recommendations. The focus of treatment has changed from LDL-C and TC goals to stratifying patients to either high-intensity or moderate-intensity statin therapy, based on their comorbidities and risk of atherosclerotic CV disease (ASCVD).¹ Primary prevention of CV disease has been proposed for patients with diabetes mellitus aged 40 to 75 years, familial hypercholesterolemia (LDL-C > 190 mg/dL), and for patients with an ASCVD risk score > 7.5%. Secondary prevention has been proposed for all patients with a history of ASCVD. Among the available choices for intensive statin therapy, the 2 most potent regimens are atorvastatin (40-80 mg) and rosuvastatin (20-40 mg) daily. The ACC/AHA guideline recommends high-potency therapy with either rosuvastatin or atorvastatin with equal preference.¹

Related: New Incentives for Helping Prevent Heart Disease

Statin therapy is generally well tolerated; however, the use of statins is not without risk of adverse drug reactions (ADRs). Skeletal muscle discomfort has been reported in 4% to 10% of patients taking either atorvastatin or rosuvastatin.^4,5 Liver enzyme abnormalities are less common, having been reported in only about 2% to 3% of patients taking either atorvastatin or rosuvastatin.^4,5 Although muscle-related intolerance and liver enzyme abnormalities are considered to be class effects, research speculates that the therapeutic and safety effects of statins may differ, based on tissue solubility.^2,6 Variability in the myotoxic and hepatotoxic effects of statins has been attributed to differences in tissue solubility, hypothesizing that lipophilic statins are more easily taken up into myocytes and hepatocytes, resulting in an increase in toxic effects.^2,6

This study assessed differences in therapeutic and safety endpoints resulting from the recent interchange from rosuvastatin to atorvastatin within the North Florida/South Georgia Veterans Health System (NF/SGVHS). Although both these agents are high-potency HMG Co-A reductase inhibitors and share a mechanism of action, they have pharmacokinetic differences, including a key difference in tissue solubility (Table 1).

With the availability of low-cost generic atorvastatin in early 2012, the VA Medical Advisory Panel and Pharmacy Benefits Management (VA MAP/PBM) added atorvastatin to the VA National Formulary as the preferred high-potency statin.^7,8 Before October 2012, rosuvastatin had been the preferred high-potency statin within the VA. With support from VA MAP/PBM leadership, NF/SGVHS instituted an interchange from rosuvastatin to atorvastatin for cost-savings purposes.⁹

Before the interchange, the records of patients were reviewed to determine whether justification existed for continued use of rosuvastatin. Patients were converted to atorvastatin if deemed appropriate and received education and consultation through direct patient contact or a letter regarding the interchange. Justifications for continued use of rosuvastatin included documentation of an atorvastatin ADR, active liver disease, or patients taking cyclosporine or certain protease inhibitors.⁸

The objective of this retrospective evaluation was to assess the efficacy and safety of the interchange from rosuvastatin to atorvastatin within NF/SGVHS. The results of this review are helpful to confirm the efficacy and safety of the interchange and identify any differences in efficacy and safety that may have occurred as a result of the interchange to a different high-potency statin. For this review, statin efficacy was assessed via review of pre- and postinterchange lipid panel values, assessing for a significant difference between equipotent atorvastatin and rosuvastatin therapy. Similarly, safety was assessed by analysis of pre- and postinterchange liver enzyme panels and assessing for significant differences as a result of the interchange.

Methods

The therapeutic interchange was conducted within the NF/SGVHS, which provides patient care at hospitals in Gainesville, Florida, and Lake City, Florida, and 11 outpatient clinics located throughout North Florida and South Georgia. Like other VA facilities, NF/SGVHS uses Computerized Patient Records System (CPRS) to electronically integrate all clinical patient information, including medical progress notes, consults, admission and discharge summaries, allergies and ADRs, patient problem lists (diagnoses), vital signs, medication orders, and laboratory test results. Approval to conduct this study was granted by the University of Florida Investigational Review Board and the Research and Development Committee at NF/SGVHS.

Therapeutic Interchange Process

Interchange from rosuvastatin to atorvastatin was expected to provide about $643,000 annually in drug cost savings to NF/SGVHS while providing equivalent therapy. The cost for a 30-day supply of rosuvastatin was about $22.56 at the time of interchange, and a 30-day supply of generic atorvastatin was $1.77.The interchange from rosuvastatin to atorvastatin was approved by the Pharmacy and Therapeutics Committee Meeting on August 8, 2012, and began shortly thereafter. Interchanges were halted temporarily in November 2012 due to a shortage of manufacturer supply, but the process fully resumed in January 2013 once the drug shortage resolved. Direction was provided to VA facilities by a guidance letter issued through PBM leadership.⁸ The interchange used a standard interchange guide to complete the process (Table 2).

Posttherapeutic Interchange Analysis

Researchers conducted an internal pharmacy computerizedprescription records search to identify VA outpatients who were converted from rosuvastatin to atorvastatin from February 1, 2012, to August 1, 2013. A total of 202 patients were randomly selected and included in this retrospective chart review. Investigators analyzed data points for safety and efficacy, including liver function tests (LFTs), lipid panels, and ADR reports. This information was obtained from laboratory data, vital signs, allergy information, ADR data, and progress notes using CPRS. Two sets of laboratory data were obtained for research purposes, the most recent laboratory values pre-interchange and the most recent laboratory values postinterchange to atorvastatin.

Investigators determined whether patients were converted to an equivalent dose of atorvastatin through an assessment of the most recent dosage of rosuvastatin before the interchange and the dosage of atorvastatin postinterchange. Researchers also analyzed refill history and interacting medications to assess possible confounding factors. Medication adherence was assessed via refill history. Medication adherence was defined as a medication possession ratio of at least 70%, which correlated to receipt of 3 or more 90-day supplies in the year prior to interchange. The above data were collected via retrospective chart review and entered into a spreadsheet.

Statistics

All identifying information was removed from the data set prior to statistical analysis. The data analysis for this project was generated using SAS/STAT software, version 9.3 of the SAS System for Linux x64 (SAS Institute, Cary, North Carolina).

Researchers assumed an equal variance in preinterchange and postinterchange lipid and liver panel values. Preinterchange and postinterchange lipid values (LDL-C, HDL-C, total cholesterol [TC], and triglycerides [TGs]) and liver values (aspartate aminotransferase [AST], alanine aminotransferase [ALT], alkaline phosphatase [ALP], and creatinine phosphokinase [CPK]) were analyzed by paired t test. All values were reported as mean (SD), and significance was defined as P < .05.

Results

More than 6,000 veterans within the NF/SGVHS were identified as eligible for the interchange. Of those who were converted, 202 patient records were randomly selected and reviewed. Patient population characteristics are summarized in Table 3. Most patients were aged > 65 years (61.4%) with an average body mass index (BMI) of 32.4. Most patients were converted to the correct corresponding dose of atorvastatin (82.7%) and achieved adherence with statin therapy (84.2%). There was no difference in pre- and postinterchange adherence detected as a result of this review.

Related: New Guideline on Dyslipidemia: Less Is More

Adverse drug reactions were documented in 16 cases, accounting for 8% of the study population. The most commonly reported ADR was myalgia or arthralgia, which was found in 10 cases (5%). Other ADRs identified in this retrospective review included treatment failure, nausea, vomiting, abdominal discomfort, abnormal liver enzymes, nasopharyngitis, and pruritis (Table 4). Of note, treatment failure was determined on a case-by-case basis but was generally defined as a failure to reach LDL-C goal (< 100 mg/dL or < 70 mg/dL), despite titration of atorvastatin. Interacting medications were identified in 30.7% of patients; however, no reported ADRs were associated with interacting medications. The most common drug interaction was concomitant niacin, followed by antiarrhythmics (ie, amiodarone, diltiazem, etc), and fibrates (ie, gemfibrozil, fenofibrate). All potentially interacting medications identified in this retrospective chart review are compiled in Table 5.

No significant difference between mean pre- and postinterchange lipid panel values was identified in this retrospective chart review (Table 6). In addition, no significant difference was detected in pre- and post-interchange AST, ALT, and CPK values (Table 7). However, a statistically significant increase in ALP was detected, with a mean ALP of 73.33 IU/L prior to interchange and 83.64 IU/L postinterchange (P < .0001).

Discussion

The goal of this retrospective observation was to ensure that safety and efficacy were not compromised as a result of this cost-saving therapeutic interchange. No differences in liver enzymes (safety) and lipid control (effectiveness) were observed in this study. There were no statistically significant changes to the lipid panel or liver panel detected with the exception of ALP. The reason for this statistically significant increase is unknown; however, it may support the hypothesis of variation in hepatocellular effects between the statins due to lipophilic properties.^3,6 In general, liver enzymes can be affected by extrahepatic functions. Serum ALP and other liver enzymes can be affected by bone disease, abdominal adiposity, alcoholism, and other concomitant diseases.¹⁰ No comorbid conditions were assessed, thus differences in liver enzymes may not be fully attributable to statin therapy. This retrospective review found no clinically significant effect correlated with the increase in ALP.

The results of this analysis are congruent with similar therapeutic interchange studies, which resulted in cost savings without compromising safety or efficacy.^9,11 Unlike other therapeutic interchange studies, this study analyzed both safety and lipid-lowering efficacy outcomes, instead of focusing solely on changes in LDL-C lowering, total cost savings, and/or adherence.^12,13 By including the entire lipid panel and liver panel into the review, this study conducted a more inclusive review of interchangeability with statins, addressing issues such as HDL-C lowering, TG changes, and liver enzyme fluctuation on conversion. There had not been a sufficient time to assess efficacy in terms of CV outcomes.

Related: Poor Outcomes for African Americans in Cardiac Rehabilitation

Two adverse events alluded to therapeutic failure as a reason for discontinuing atorvastatin. In the previous ATP III lipid guidelines, therapeutic failure was achieved when patients did not reach their LDL-C goals despite appropriate titration of statin therapy.² However, the ACC/AHA lipid guidelines have done away with lipid goals as a measurement of treatment therapy, focusing rather on evidence-based high- or moderate-intensity statin therapy that has been proven in clinical trials to reduce mortality and CV events.¹ Although measurement of efficacy via lipid panel values is no longer a guideline recommendation, the results of this chart review have shown no difference in lipid values as a result of the interchange, confirming the interchangeability of rosuvastatin and atorvastatin at their equivalent doses.

Limitations

The interchange of rosuvastatin to atorvastatin was a policy change affecting all patients within the NF/SGVHS. In order to reflect true population data and more accurately predict the effects of such a policy change, this study used intention-to-treat analysis, including all patients, even patients who were found to be nonadherent. This study is also limited by sample size (N = 202). Additionally, the generalizability of these findings may be limited. The study population was mostly males aged > 65 years with an average BMI of 32.4. Researchers did not compile comorbidity, race, or concomitant medication data. Additionally, the duration of statin therapy prior to laboratory value collection was undefined.

A retrospective chart review lends itself to limitations in data collection. Medication adherence is a factor that is assumed to have a significant effect on the results of this interchange. In this review, adherence was assessed via refill history. Researchers were unable to confirm actual consumption of the medication.

Additionally, researchers did not analyze comorbid conditions, which may have had an effect on lipid panel and liver panel values. For those veterans who discontinued atorvastatin therapy, the reason for discontinuation was often not documented. Thus, researchers were unable to assess reasons for discontinuation.

Conclusion

The results generated from a review of the therapeutic exchange of rosuvastatin to atorvastatin within a veteran population affirm that the interchange was not associated with any differences in safety or lipid control, but did result in significant drug cost savings. This study provides support for health care systems considering therapeutic interchange with high-intensity statins safely and effectively.

Acknowledgements
This material is the result of work supported with resources and the use of facilities at the North Florida/South Georgia Veterans Health System in Gainesville, Florida. The authors would like to acknowledge Kim Hoang, PharmD, for her contributions to this project.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer


The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Patients with known cardiovascular (CV) disease are at greater risk for CV events.¹ Hydroxymethylglutaryl-CoA (HMG Co-A) reductase inhibitors, or statins, have been shown to reduce CV events and to reduce all-cause mortality.^1,2 Thus, these agents should be a standard approach to secondary prevention of CV events.^1,2 Although the main function of statins is to lower total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) levels, trials have shown that other lipid-lowering agents have reduced the incidence of CV events but have failed to show any difference in mortality.¹ Thus, the therapeutic effects of statins may be a result of “pleiotropic” effects in addition to a reduction in LDL-C.³

As a result, prescribing practices and professional society guidelines have deferred to statins as a first-line choice for lipid-lowering therapy.¹ Although each statin varies in its ability to lower LDL-C and inhibit HMG Co-A reductase, as a class, statins have been proven to be safe and efficacious in reducing LDL-C, decreasing risk of coronary artery disease, and decreasing mortality.

The 2013 American College of Cardiology (ACC) and American Heart Association (AHA) guideline on the treatment of blood cholesterol to reduce CV risk in adults resulted in a major shift in clinical practice recommendations. The focus of treatment has changed from LDL-C and TC goals to stratifying patients to either high-intensity or moderate-intensity statin therapy, based on their comorbidities and risk of atherosclerotic CV disease (ASCVD).¹ Primary prevention of CV disease has been proposed for patients with diabetes mellitus aged 40 to 75 years, familial hypercholesterolemia (LDL-C > 190 mg/dL), and for patients with an ASCVD risk score > 7.5%. Secondary prevention has been proposed for all patients with a history of ASCVD. Among the available choices for intensive statin therapy, the 2 most potent regimens are atorvastatin (40-80 mg) and rosuvastatin (20-40 mg) daily. The ACC/AHA guideline recommends high-potency therapy with either rosuvastatin or atorvastatin with equal preference.¹

Related: New Incentives for Helping Prevent Heart Disease

Statin therapy is generally well tolerated; however, the use of statins is not without risk of adverse drug reactions (ADRs). Skeletal muscle discomfort has been reported in 4% to 10% of patients taking either atorvastatin or rosuvastatin.^4,5 Liver enzyme abnormalities are less common, having been reported in only about 2% to 3% of patients taking either atorvastatin or rosuvastatin.^4,5 Although muscle-related intolerance and liver enzyme abnormalities are considered to be class effects, research speculates that the therapeutic and safety effects of statins may differ, based on tissue solubility.^2,6 Variability in the myotoxic and hepatotoxic effects of statins has been attributed to differences in tissue solubility, hypothesizing that lipophilic statins are more easily taken up into myocytes and hepatocytes, resulting in an increase in toxic effects.^2,6

This study assessed differences in therapeutic and safety endpoints resulting from the recent interchange from rosuvastatin to atorvastatin within the North Florida/South Georgia Veterans Health System (NF/SGVHS). Although both these agents are high-potency HMG Co-A reductase inhibitors and share a mechanism of action, they have pharmacokinetic differences, including a key difference in tissue solubility (Table 1).

With the availability of low-cost generic atorvastatin in early 2012, the VA Medical Advisory Panel and Pharmacy Benefits Management (VA MAP/PBM) added atorvastatin to the VA National Formulary as the preferred high-potency statin.^7,8 Before October 2012, rosuvastatin had been the preferred high-potency statin within the VA. With support from VA MAP/PBM leadership, NF/SGVHS instituted an interchange from rosuvastatin to atorvastatin for cost-savings purposes.⁹

Before the interchange, the records of patients were reviewed to determine whether justification existed for continued use of rosuvastatin. Patients were converted to atorvastatin if deemed appropriate and received education and consultation through direct patient contact or a letter regarding the interchange. Justifications for continued use of rosuvastatin included documentation of an atorvastatin ADR, active liver disease, or patients taking cyclosporine or certain protease inhibitors.⁸

The objective of this retrospective evaluation was to assess the efficacy and safety of the interchange from rosuvastatin to atorvastatin within NF/SGVHS. The results of this review are helpful to confirm the efficacy and safety of the interchange and identify any differences in efficacy and safety that may have occurred as a result of the interchange to a different high-potency statin. For this review, statin efficacy was assessed via review of pre- and postinterchange lipid panel values, assessing for a significant difference between equipotent atorvastatin and rosuvastatin therapy. Similarly, safety was assessed by analysis of pre- and postinterchange liver enzyme panels and assessing for significant differences as a result of the interchange.

Methods

The therapeutic interchange was conducted within the NF/SGVHS, which provides patient care at hospitals in Gainesville, Florida, and Lake City, Florida, and 11 outpatient clinics located throughout North Florida and South Georgia. Like other VA facilities, NF/SGVHS uses Computerized Patient Records System (CPRS) to electronically integrate all clinical patient information, including medical progress notes, consults, admission and discharge summaries, allergies and ADRs, patient problem lists (diagnoses), vital signs, medication orders, and laboratory test results. Approval to conduct this study was granted by the University of Florida Investigational Review Board and the Research and Development Committee at NF/SGVHS.

Therapeutic Interchange Process

Interchange from rosuvastatin to atorvastatin was expected to provide about $643,000 annually in drug cost savings to NF/SGVHS while providing equivalent therapy. The cost for a 30-day supply of rosuvastatin was about $22.56 at the time of interchange, and a 30-day supply of generic atorvastatin was $1.77.The interchange from rosuvastatin to atorvastatin was approved by the Pharmacy and Therapeutics Committee Meeting on August 8, 2012, and began shortly thereafter. Interchanges were halted temporarily in November 2012 due to a shortage of manufacturer supply, but the process fully resumed in January 2013 once the drug shortage resolved. Direction was provided to VA facilities by a guidance letter issued through PBM leadership.⁸ The interchange used a standard interchange guide to complete the process (Table 2).

Posttherapeutic Interchange Analysis

Researchers conducted an internal pharmacy computerizedprescription records search to identify VA outpatients who were converted from rosuvastatin to atorvastatin from February 1, 2012, to August 1, 2013. A total of 202 patients were randomly selected and included in this retrospective chart review. Investigators analyzed data points for safety and efficacy, including liver function tests (LFTs), lipid panels, and ADR reports. This information was obtained from laboratory data, vital signs, allergy information, ADR data, and progress notes using CPRS. Two sets of laboratory data were obtained for research purposes, the most recent laboratory values pre-interchange and the most recent laboratory values postinterchange to atorvastatin.

Investigators determined whether patients were converted to an equivalent dose of atorvastatin through an assessment of the most recent dosage of rosuvastatin before the interchange and the dosage of atorvastatin postinterchange. Researchers also analyzed refill history and interacting medications to assess possible confounding factors. Medication adherence was assessed via refill history. Medication adherence was defined as a medication possession ratio of at least 70%, which correlated to receipt of 3 or more 90-day supplies in the year prior to interchange. The above data were collected via retrospective chart review and entered into a spreadsheet.

Statistics

All identifying information was removed from the data set prior to statistical analysis. The data analysis for this project was generated using SAS/STAT software, version 9.3 of the SAS System for Linux x64 (SAS Institute, Cary, North Carolina).

Researchers assumed an equal variance in preinterchange and postinterchange lipid and liver panel values. Preinterchange and postinterchange lipid values (LDL-C, HDL-C, total cholesterol [TC], and triglycerides [TGs]) and liver values (aspartate aminotransferase [AST], alanine aminotransferase [ALT], alkaline phosphatase [ALP], and creatinine phosphokinase [CPK]) were analyzed by paired t test. All values were reported as mean (SD), and significance was defined as P < .05.

Results

More than 6,000 veterans within the NF/SGVHS were identified as eligible for the interchange. Of those who were converted, 202 patient records were randomly selected and reviewed. Patient population characteristics are summarized in Table 3. Most patients were aged > 65 years (61.4%) with an average body mass index (BMI) of 32.4. Most patients were converted to the correct corresponding dose of atorvastatin (82.7%) and achieved adherence with statin therapy (84.2%). There was no difference in pre- and postinterchange adherence detected as a result of this review.

Related: New Guideline on Dyslipidemia: Less Is More

Adverse drug reactions were documented in 16 cases, accounting for 8% of the study population. The most commonly reported ADR was myalgia or arthralgia, which was found in 10 cases (5%). Other ADRs identified in this retrospective review included treatment failure, nausea, vomiting, abdominal discomfort, abnormal liver enzymes, nasopharyngitis, and pruritis (Table 4). Of note, treatment failure was determined on a case-by-case basis but was generally defined as a failure to reach LDL-C goal (< 100 mg/dL or < 70 mg/dL), despite titration of atorvastatin. Interacting medications were identified in 30.7% of patients; however, no reported ADRs were associated with interacting medications. The most common drug interaction was concomitant niacin, followed by antiarrhythmics (ie, amiodarone, diltiazem, etc), and fibrates (ie, gemfibrozil, fenofibrate). All potentially interacting medications identified in this retrospective chart review are compiled in Table 5.

No significant difference between mean pre- and postinterchange lipid panel values was identified in this retrospective chart review (Table 6). In addition, no significant difference was detected in pre- and post-interchange AST, ALT, and CPK values (Table 7). However, a statistically significant increase in ALP was detected, with a mean ALP of 73.33 IU/L prior to interchange and 83.64 IU/L postinterchange (P < .0001).

Discussion

The goal of this retrospective observation was to ensure that safety and efficacy were not compromised as a result of this cost-saving therapeutic interchange. No differences in liver enzymes (safety) and lipid control (effectiveness) were observed in this study. There were no statistically significant changes to the lipid panel or liver panel detected with the exception of ALP. The reason for this statistically significant increase is unknown; however, it may support the hypothesis of variation in hepatocellular effects between the statins due to lipophilic properties.^3,6 In general, liver enzymes can be affected by extrahepatic functions. Serum ALP and other liver enzymes can be affected by bone disease, abdominal adiposity, alcoholism, and other concomitant diseases.¹⁰ No comorbid conditions were assessed, thus differences in liver enzymes may not be fully attributable to statin therapy. This retrospective review found no clinically significant effect correlated with the increase in ALP.

The results of this analysis are congruent with similar therapeutic interchange studies, which resulted in cost savings without compromising safety or efficacy.^9,11 Unlike other therapeutic interchange studies, this study analyzed both safety and lipid-lowering efficacy outcomes, instead of focusing solely on changes in LDL-C lowering, total cost savings, and/or adherence.^12,13 By including the entire lipid panel and liver panel into the review, this study conducted a more inclusive review of interchangeability with statins, addressing issues such as HDL-C lowering, TG changes, and liver enzyme fluctuation on conversion. There had not been a sufficient time to assess efficacy in terms of CV outcomes.

Related: Poor Outcomes for African Americans in Cardiac Rehabilitation

Two adverse events alluded to therapeutic failure as a reason for discontinuing atorvastatin. In the previous ATP III lipid guidelines, therapeutic failure was achieved when patients did not reach their LDL-C goals despite appropriate titration of statin therapy.² However, the ACC/AHA lipid guidelines have done away with lipid goals as a measurement of treatment therapy, focusing rather on evidence-based high- or moderate-intensity statin therapy that has been proven in clinical trials to reduce mortality and CV events.¹ Although measurement of efficacy via lipid panel values is no longer a guideline recommendation, the results of this chart review have shown no difference in lipid values as a result of the interchange, confirming the interchangeability of rosuvastatin and atorvastatin at their equivalent doses.

Limitations

The interchange of rosuvastatin to atorvastatin was a policy change affecting all patients within the NF/SGVHS. In order to reflect true population data and more accurately predict the effects of such a policy change, this study used intention-to-treat analysis, including all patients, even patients who were found to be nonadherent. This study is also limited by sample size (N = 202). Additionally, the generalizability of these findings may be limited. The study population was mostly males aged > 65 years with an average BMI of 32.4. Researchers did not compile comorbidity, race, or concomitant medication data. Additionally, the duration of statin therapy prior to laboratory value collection was undefined.

A retrospective chart review lends itself to limitations in data collection. Medication adherence is a factor that is assumed to have a significant effect on the results of this interchange. In this review, adherence was assessed via refill history. Researchers were unable to confirm actual consumption of the medication.

Additionally, researchers did not analyze comorbid conditions, which may have had an effect on lipid panel and liver panel values. For those veterans who discontinued atorvastatin therapy, the reason for discontinuation was often not documented. Thus, researchers were unable to assess reasons for discontinuation.

Conclusion

The results generated from a review of the therapeutic exchange of rosuvastatin to atorvastatin within a veteran population affirm that the interchange was not associated with any differences in safety or lipid control, but did result in significant drug cost savings. This study provides support for health care systems considering therapeutic interchange with high-intensity statins safely and effectively.

Acknowledgements
This material is the result of work supported with resources and the use of facilities at the North Florida/South Georgia Veterans Health System in Gainesville, Florida. The authors would like to acknowledge Kim Hoang, PharmD, for her contributions to this project.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer


The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Due to the growing number of drug shortages and increasing cost of medications, large-scale formulary conversions are becoming more common and necessary for pharmacies to stay within their budget allocations. Statins are widely recognized as first-line therapy for cholesterol lowering and have been proven to reduce cardiovascular morbidity and mortality.^1-5 In addition to statin therapy, weight loss and lifestyle changes are often necessary to meet optimum low-density lipoprotein cholesterol (LDL-C) goals.⁶ According to the Adult Treatment Panel III (ATP III) guidelines, the optimum LDL-C for each patient varies based on the presence of coronary artery disease (CAD), CAD risk equivalents, and other risk factors.⁷ Patients with CAD or CAD risk equivalents have an LDL-C goal of < 100 mg/dL. Those with multiple risk factors have an LDL-C goal of < 130 mg/dL, and those with 0 to 1 risk factors have an LDL-C goal of < 160 mg/dL.^1-3,7,8

Numerous trials have compared the safety and efficacy of the 3-hydroxy-3-methylglutaryl-coenzyme A inhibitors, stating that rosuvastatin 5 to 10 mg per day is equivalent to 20 mg per day of atorvastatin in terms of its ability to lower LDL-C levels.^7,9-14 The LUNAR (Limiting Under treatment of lipids in Acute coronary syndrome with Rosuvastatin) study compared the efficacy of rosuvastatin with that of atorvastatin in decreasing LDL-C in patients with acute coronary syndrome.⁸ Rosuvastatin 40 mg was significantly more successful in lowering LDL-C and increasing high-density lipoprotein cholesterol (HDL-C) compared with atorvastatin 80-mg daily therapy.

In October 2012, the national VA Pharmacy Benefits Management (PBM) Services released guidance regarding the conversion from rosuvastatin to atorvastatin, including the dosing conversion (Table 1), stating that rosuvastatin 5 mg daily should be considered equivalent and converted to atorvastatin 20 mg daily. In general, adverse events (AEs), such as increased liver enzymes, myopathies, and increased creatinine phosphokinase (CPK), are considered a class effect of the statins.^13,14

Related: Statins showed no benefit in reducing risk of recurrent VTE

The recent 2013 American College of Cardiology/American Heart Association (ACC/AHA) Blood Cholesterol Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults has identified a large number of patients as candidates for high-intensity statins, which the authors defined as atorvastatin and rosuvastatin.⁴ These guidelines do not recommend LDL-C goals and instead use a risk calculator to determine which intensity of statin therapy is appropriate for certain patients. This change in practice will lead to a higher volume of high-intensity statin prescriptions and higher drug costs for some medical centers. Given the volume of prescriptions and the increased use of large-scale formulary conversions to reduce costs, more research is warranted to ensure equivalent dosing. With more data available and equivalent dosing defined, pharmacists may be better able to improve clinical results in patients that are included in these large-scale formulary conversions.

Methods

A retrospective chart review was performed on all LDL-C levels in patients receiving atorvastatin therapy due to the formulary conversion from rosuvastatin to atorvastatin at the Huntington VAMC (HVAMC) in Huntington, West Virginia, per the guidance published by the national VA PBM Services. The number of patients not at their LDL-C goal (as defined by ATP III guidelines) as a result of atorvastatin therapy was determined. Furthermore, AEs due to atorvastatin therapy such as increased liver enzymes, myopathy, and increased CPK were identified and analyzed.

The primary endpoint of this study focused on the rate of treatment failure in LDL-C reduction with atorvastatin treatment in patients previously at their LDL-C goal, as defined by the ATP III guidelines, with rosuvastatin therapy. Secondary endpoints included rate of AEs due to atorvastatin therapy, percentage increase in CPK and liver enzymes as a result of atorvastatin therapy, and percentage LDL-C, HDL-C, total cholesterol (TC), and triglyceride (TGs) changes since conversion from rosuvastatin to atorvastatin.

Patients were included in the study if they were aged 18 to 89 years, previously at LDL-C goal with rosuvastatin therapy for at least 3 months, had never previously received atorvastatin, were converted to atorvastatin therapy as a result of a large-scale formulary conversion at HVAMC, and had a fasting lipid panel completed 1 to 6 months after conversion. Patients were excluded from the study if they received other lipid-lowering medications (eg, bile acid sequestrants, fibrates, niacin, or ezetimibe) in the 12 months before or after receiving statin therapy, had previously documented AEs (eg, myopathy, increased liver enzymes, increased CPK as a result of statin therapy, or history of known homozygous familial hypercholesterolemia) current active liver disease (ALT > 2x ULN [upper limit of normal]), unexplained CPK ≥ 3x ULN, serum creatinine (SCr) > 2 mg/dL, or history of alcohol or drug abuse within the last 5 years.

Results

Three hundred twenty-three patients were identified and reviewed as converted from rosuvastatin to atorvastatin during the study period with no prior use of atorvastatin. Of the 323 charts that were reviewed, 195 patients met the study inclusion criteria and were analyzed for rate of treatment failure in terms of lipid goals and rate of AEs. Twenty of 195 patients (10.3%) were no longer at their LDL-C goal after conversion from rosuvastatin to atorvastatin. Of those 195 patients, 29 (14.9%) experienced an adverse drug reaction (ADR) as a result of atorvastatin treatment that was severe enough to result in discontinuation of the drug and switching the patient back to the originally prescribed dose of rosuvastatin. Figure 1 illustrates the number of patients and documented atorvastatin ADRs. The most common ADR documented to atorvastatin was myalgias (8.2%).

The average change in lipid levels was calculated and atorvastatin therapy was found to result in clinically insignificant changes to the lipid panel (Table 2). A 2-tailed paired t test was used to assess the statistical significance of these changes. Atorvastatin therapy resulted in an average decrease of LDL-C by 5.0 mg/dL (P < .01) in comparison to previous therapy with equivalent rosuvastatin dose. Other noted changes to lipid profile after formulary conversion included TG reduction by 2 mg/dL (P = .69), TC increased by 0.58 mg/dL (P = .80), and HDL-C reduction by 4.66 mg/dL (P < .01).

Although the decrease in LDL-C and HDL-C as a result of the formulary change was found to be statistically significant, they are not thought to result in a clinical difference. Clinically and statistically insignificant changes in liver enzymes and CPK were also discovered as a result of atorvastatin therapy conversion (Table 3). Atorvastatin therapy resulted in an averaged decrease of aspartate aminotransferase by 2.2 IU/L (P = .19) and an increase in alanine aminotransferase by 1.4 IU/L (P = .47). Average change in CPK was -6 IU/L (P = 89).

Related: New Guideline on Dyslipidemia: Less Is More

Discussion

Lipid levels were found to be mostly unchanged and remained at therapy goals, demonstrating use and appropriate equivalent dosing of rosuvastatin and atorvastatin following the formulary conversion defined in Table 1. Documented ADRs were minimal, indicating the ingredient conversion was well tolerated overall by our patients. Following the release of the 2013ACC/AHA guidelines, many patients in the VA required treatment with high-potency statins such as rosuvastatin and atorvastatin. Given the volume of statin prescriptions in the VA and the significant potential for providing the most cost-efficient lipid therapy (Table 4), the formulary conversion from rosuvastatin to atorvastatin was warranted.

Limitations

There are several limitations for extrapolating results from this study to the general population. Due to the retrospective design of the study, no formal assessment of adherence was conducted. A prospective trial, with a researcher monitoring refills and tablet counts would be more accurate to ensure patients adherence with statin therapy.

Many of the patients that were included in the formulary conversion did not have follow-up laboratory work at the time of this study and were therefore excluded, leading to a smaller study population. A larger study population with a similar study design may be able to detect more significant correlations.

There are also several potential complications in regard to formulary conversions, including the inability to ensure the exact period that the patient switched therapies. During the conversion of rosuvastatin to atorvastatin at HVAMC, an attempt was made to minimize this confounder by converting patients on request for a refill of their rosuvastatin therapy. Last, all 185 patients that were studied were male; therefore, it is difficult to extrapolate these results for female patients.

Conclusions

This study shows that the conversion of patients from rosuvastatin therapy to atorvastatin was effective when it targeted a specific LDL-C goal or specific reduction in LDL-C. A similar conversion would likely lead to lower drug costs for many other health systems. Additional studies will be necessary given the recent changes in the national lipid guidelines. Furthermore, studies will be needed to assess concrete clinical endpoints (cardiovascular mortality, all-cause mortality, and cardiac events).

Related: Poor Outcomes for African Americans in Cardiac Rehabilitation

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Due to the growing number of drug shortages and increasing cost of medications, large-scale formulary conversions are becoming more common and necessary for pharmacies to stay within their budget allocations. Statins are widely recognized as first-line therapy for cholesterol lowering and have been proven to reduce cardiovascular morbidity and mortality.^1-5 In addition to statin therapy, weight loss and lifestyle changes are often necessary to meet optimum low-density lipoprotein cholesterol (LDL-C) goals.⁶ According to the Adult Treatment Panel III (ATP III) guidelines, the optimum LDL-C for each patient varies based on the presence of coronary artery disease (CAD), CAD risk equivalents, and other risk factors.⁷ Patients with CAD or CAD risk equivalents have an LDL-C goal of < 100 mg/dL. Those with multiple risk factors have an LDL-C goal of < 130 mg/dL, and those with 0 to 1 risk factors have an LDL-C goal of < 160 mg/dL.^1-3,7,8

Numerous trials have compared the safety and efficacy of the 3-hydroxy-3-methylglutaryl-coenzyme A inhibitors, stating that rosuvastatin 5 to 10 mg per day is equivalent to 20 mg per day of atorvastatin in terms of its ability to lower LDL-C levels.^7,9-14 The LUNAR (Limiting Under treatment of lipids in Acute coronary syndrome with Rosuvastatin) study compared the efficacy of rosuvastatin with that of atorvastatin in decreasing LDL-C in patients with acute coronary syndrome.⁸ Rosuvastatin 40 mg was significantly more successful in lowering LDL-C and increasing high-density lipoprotein cholesterol (HDL-C) compared with atorvastatin 80-mg daily therapy.

In October 2012, the national VA Pharmacy Benefits Management (PBM) Services released guidance regarding the conversion from rosuvastatin to atorvastatin, including the dosing conversion (Table 1), stating that rosuvastatin 5 mg daily should be considered equivalent and converted to atorvastatin 20 mg daily. In general, adverse events (AEs), such as increased liver enzymes, myopathies, and increased creatinine phosphokinase (CPK), are considered a class effect of the statins.^13,14

Related: Statins showed no benefit in reducing risk of recurrent VTE

The recent 2013 American College of Cardiology/American Heart Association (ACC/AHA) Blood Cholesterol Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults has identified a large number of patients as candidates for high-intensity statins, which the authors defined as atorvastatin and rosuvastatin.⁴ These guidelines do not recommend LDL-C goals and instead use a risk calculator to determine which intensity of statin therapy is appropriate for certain patients. This change in practice will lead to a higher volume of high-intensity statin prescriptions and higher drug costs for some medical centers. Given the volume of prescriptions and the increased use of large-scale formulary conversions to reduce costs, more research is warranted to ensure equivalent dosing. With more data available and equivalent dosing defined, pharmacists may be better able to improve clinical results in patients that are included in these large-scale formulary conversions.

Methods

A retrospective chart review was performed on all LDL-C levels in patients receiving atorvastatin therapy due to the formulary conversion from rosuvastatin to atorvastatin at the Huntington VAMC (HVAMC) in Huntington, West Virginia, per the guidance published by the national VA PBM Services. The number of patients not at their LDL-C goal (as defined by ATP III guidelines) as a result of atorvastatin therapy was determined. Furthermore, AEs due to atorvastatin therapy such as increased liver enzymes, myopathy, and increased CPK were identified and analyzed.

The primary endpoint of this study focused on the rate of treatment failure in LDL-C reduction with atorvastatin treatment in patients previously at their LDL-C goal, as defined by the ATP III guidelines, with rosuvastatin therapy. Secondary endpoints included rate of AEs due to atorvastatin therapy, percentage increase in CPK and liver enzymes as a result of atorvastatin therapy, and percentage LDL-C, HDL-C, total cholesterol (TC), and triglyceride (TGs) changes since conversion from rosuvastatin to atorvastatin.

Patients were included in the study if they were aged 18 to 89 years, previously at LDL-C goal with rosuvastatin therapy for at least 3 months, had never previously received atorvastatin, were converted to atorvastatin therapy as a result of a large-scale formulary conversion at HVAMC, and had a fasting lipid panel completed 1 to 6 months after conversion. Patients were excluded from the study if they received other lipid-lowering medications (eg, bile acid sequestrants, fibrates, niacin, or ezetimibe) in the 12 months before or after receiving statin therapy, had previously documented AEs (eg, myopathy, increased liver enzymes, increased CPK as a result of statin therapy, or history of known homozygous familial hypercholesterolemia) current active liver disease (ALT > 2x ULN [upper limit of normal]), unexplained CPK ≥ 3x ULN, serum creatinine (SCr) > 2 mg/dL, or history of alcohol or drug abuse within the last 5 years.

Results

Three hundred twenty-three patients were identified and reviewed as converted from rosuvastatin to atorvastatin during the study period with no prior use of atorvastatin. Of the 323 charts that were reviewed, 195 patients met the study inclusion criteria and were analyzed for rate of treatment failure in terms of lipid goals and rate of AEs. Twenty of 195 patients (10.3%) were no longer at their LDL-C goal after conversion from rosuvastatin to atorvastatin. Of those 195 patients, 29 (14.9%) experienced an adverse drug reaction (ADR) as a result of atorvastatin treatment that was severe enough to result in discontinuation of the drug and switching the patient back to the originally prescribed dose of rosuvastatin. Figure 1 illustrates the number of patients and documented atorvastatin ADRs. The most common ADR documented to atorvastatin was myalgias (8.2%).

The average change in lipid levels was calculated and atorvastatin therapy was found to result in clinically insignificant changes to the lipid panel (Table 2). A 2-tailed paired t test was used to assess the statistical significance of these changes. Atorvastatin therapy resulted in an average decrease of LDL-C by 5.0 mg/dL (P < .01) in comparison to previous therapy with equivalent rosuvastatin dose. Other noted changes to lipid profile after formulary conversion included TG reduction by 2 mg/dL (P = .69), TC increased by 0.58 mg/dL (P = .80), and HDL-C reduction by 4.66 mg/dL (P < .01).

Although the decrease in LDL-C and HDL-C as a result of the formulary change was found to be statistically significant, they are not thought to result in a clinical difference. Clinically and statistically insignificant changes in liver enzymes and CPK were also discovered as a result of atorvastatin therapy conversion (Table 3). Atorvastatin therapy resulted in an averaged decrease of aspartate aminotransferase by 2.2 IU/L (P = .19) and an increase in alanine aminotransferase by 1.4 IU/L (P = .47). Average change in CPK was -6 IU/L (P = 89).

Related: New Guideline on Dyslipidemia: Less Is More

Discussion

Lipid levels were found to be mostly unchanged and remained at therapy goals, demonstrating use and appropriate equivalent dosing of rosuvastatin and atorvastatin following the formulary conversion defined in Table 1. Documented ADRs were minimal, indicating the ingredient conversion was well tolerated overall by our patients. Following the release of the 2013ACC/AHA guidelines, many patients in the VA required treatment with high-potency statins such as rosuvastatin and atorvastatin. Given the volume of statin prescriptions in the VA and the significant potential for providing the most cost-efficient lipid therapy (Table 4), the formulary conversion from rosuvastatin to atorvastatin was warranted.

Limitations

There are several limitations for extrapolating results from this study to the general population. Due to the retrospective design of the study, no formal assessment of adherence was conducted. A prospective trial, with a researcher monitoring refills and tablet counts would be more accurate to ensure patients adherence with statin therapy.

Many of the patients that were included in the formulary conversion did not have follow-up laboratory work at the time of this study and were therefore excluded, leading to a smaller study population. A larger study population with a similar study design may be able to detect more significant correlations.

There are also several potential complications in regard to formulary conversions, including the inability to ensure the exact period that the patient switched therapies. During the conversion of rosuvastatin to atorvastatin at HVAMC, an attempt was made to minimize this confounder by converting patients on request for a refill of their rosuvastatin therapy. Last, all 185 patients that were studied were male; therefore, it is difficult to extrapolate these results for female patients.

Conclusions

This study shows that the conversion of patients from rosuvastatin therapy to atorvastatin was effective when it targeted a specific LDL-C goal or specific reduction in LDL-C. A similar conversion would likely lead to lower drug costs for many other health systems. Additional studies will be necessary given the recent changes in the national lipid guidelines. Furthermore, studies will be needed to assess concrete clinical endpoints (cardiovascular mortality, all-cause mortality, and cardiac events).

Related: Poor Outcomes for African Americans in Cardiac Rehabilitation

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Due to the growing number of drug shortages and increasing cost of medications, large-scale formulary conversions are becoming more common and necessary for pharmacies to stay within their budget allocations. Statins are widely recognized as first-line therapy for cholesterol lowering and have been proven to reduce cardiovascular morbidity and mortality.^1-5 In addition to statin therapy, weight loss and lifestyle changes are often necessary to meet optimum low-density lipoprotein cholesterol (LDL-C) goals.⁶ According to the Adult Treatment Panel III (ATP III) guidelines, the optimum LDL-C for each patient varies based on the presence of coronary artery disease (CAD), CAD risk equivalents, and other risk factors.⁷ Patients with CAD or CAD risk equivalents have an LDL-C goal of < 100 mg/dL. Those with multiple risk factors have an LDL-C goal of < 130 mg/dL, and those with 0 to 1 risk factors have an LDL-C goal of < 160 mg/dL.^1-3,7,8

Numerous trials have compared the safety and efficacy of the 3-hydroxy-3-methylglutaryl-coenzyme A inhibitors, stating that rosuvastatin 5 to 10 mg per day is equivalent to 20 mg per day of atorvastatin in terms of its ability to lower LDL-C levels.^7,9-14 The LUNAR (Limiting Under treatment of lipids in Acute coronary syndrome with Rosuvastatin) study compared the efficacy of rosuvastatin with that of atorvastatin in decreasing LDL-C in patients with acute coronary syndrome.⁸ Rosuvastatin 40 mg was significantly more successful in lowering LDL-C and increasing high-density lipoprotein cholesterol (HDL-C) compared with atorvastatin 80-mg daily therapy.

In October 2012, the national VA Pharmacy Benefits Management (PBM) Services released guidance regarding the conversion from rosuvastatin to atorvastatin, including the dosing conversion (Table 1), stating that rosuvastatin 5 mg daily should be considered equivalent and converted to atorvastatin 20 mg daily. In general, adverse events (AEs), such as increased liver enzymes, myopathies, and increased creatinine phosphokinase (CPK), are considered a class effect of the statins.^13,14

Related: Statins showed no benefit in reducing risk of recurrent VTE

The recent 2013 American College of Cardiology/American Heart Association (ACC/AHA) Blood Cholesterol Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults has identified a large number of patients as candidates for high-intensity statins, which the authors defined as atorvastatin and rosuvastatin.⁴ These guidelines do not recommend LDL-C goals and instead use a risk calculator to determine which intensity of statin therapy is appropriate for certain patients. This change in practice will lead to a higher volume of high-intensity statin prescriptions and higher drug costs for some medical centers. Given the volume of prescriptions and the increased use of large-scale formulary conversions to reduce costs, more research is warranted to ensure equivalent dosing. With more data available and equivalent dosing defined, pharmacists may be better able to improve clinical results in patients that are included in these large-scale formulary conversions.

Methods

A retrospective chart review was performed on all LDL-C levels in patients receiving atorvastatin therapy due to the formulary conversion from rosuvastatin to atorvastatin at the Huntington VAMC (HVAMC) in Huntington, West Virginia, per the guidance published by the national VA PBM Services. The number of patients not at their LDL-C goal (as defined by ATP III guidelines) as a result of atorvastatin therapy was determined. Furthermore, AEs due to atorvastatin therapy such as increased liver enzymes, myopathy, and increased CPK were identified and analyzed.

The primary endpoint of this study focused on the rate of treatment failure in LDL-C reduction with atorvastatin treatment in patients previously at their LDL-C goal, as defined by the ATP III guidelines, with rosuvastatin therapy. Secondary endpoints included rate of AEs due to atorvastatin therapy, percentage increase in CPK and liver enzymes as a result of atorvastatin therapy, and percentage LDL-C, HDL-C, total cholesterol (TC), and triglyceride (TGs) changes since conversion from rosuvastatin to atorvastatin.

Patients were included in the study if they were aged 18 to 89 years, previously at LDL-C goal with rosuvastatin therapy for at least 3 months, had never previously received atorvastatin, were converted to atorvastatin therapy as a result of a large-scale formulary conversion at HVAMC, and had a fasting lipid panel completed 1 to 6 months after conversion. Patients were excluded from the study if they received other lipid-lowering medications (eg, bile acid sequestrants, fibrates, niacin, or ezetimibe) in the 12 months before or after receiving statin therapy, had previously documented AEs (eg, myopathy, increased liver enzymes, increased CPK as a result of statin therapy, or history of known homozygous familial hypercholesterolemia) current active liver disease (ALT > 2x ULN [upper limit of normal]), unexplained CPK ≥ 3x ULN, serum creatinine (SCr) > 2 mg/dL, or history of alcohol or drug abuse within the last 5 years.

Results

Three hundred twenty-three patients were identified and reviewed as converted from rosuvastatin to atorvastatin during the study period with no prior use of atorvastatin. Of the 323 charts that were reviewed, 195 patients met the study inclusion criteria and were analyzed for rate of treatment failure in terms of lipid goals and rate of AEs. Twenty of 195 patients (10.3%) were no longer at their LDL-C goal after conversion from rosuvastatin to atorvastatin. Of those 195 patients, 29 (14.9%) experienced an adverse drug reaction (ADR) as a result of atorvastatin treatment that was severe enough to result in discontinuation of the drug and switching the patient back to the originally prescribed dose of rosuvastatin. Figure 1 illustrates the number of patients and documented atorvastatin ADRs. The most common ADR documented to atorvastatin was myalgias (8.2%).

The average change in lipid levels was calculated and atorvastatin therapy was found to result in clinically insignificant changes to the lipid panel (Table 2). A 2-tailed paired t test was used to assess the statistical significance of these changes. Atorvastatin therapy resulted in an average decrease of LDL-C by 5.0 mg/dL (P < .01) in comparison to previous therapy with equivalent rosuvastatin dose. Other noted changes to lipid profile after formulary conversion included TG reduction by 2 mg/dL (P = .69), TC increased by 0.58 mg/dL (P = .80), and HDL-C reduction by 4.66 mg/dL (P < .01).

Although the decrease in LDL-C and HDL-C as a result of the formulary change was found to be statistically significant, they are not thought to result in a clinical difference. Clinically and statistically insignificant changes in liver enzymes and CPK were also discovered as a result of atorvastatin therapy conversion (Table 3). Atorvastatin therapy resulted in an averaged decrease of aspartate aminotransferase by 2.2 IU/L (P = .19) and an increase in alanine aminotransferase by 1.4 IU/L (P = .47). Average change in CPK was -6 IU/L (P = 89).

Related: New Guideline on Dyslipidemia: Less Is More

Discussion

Lipid levels were found to be mostly unchanged and remained at therapy goals, demonstrating use and appropriate equivalent dosing of rosuvastatin and atorvastatin following the formulary conversion defined in Table 1. Documented ADRs were minimal, indicating the ingredient conversion was well tolerated overall by our patients. Following the release of the 2013ACC/AHA guidelines, many patients in the VA required treatment with high-potency statins such as rosuvastatin and atorvastatin. Given the volume of statin prescriptions in the VA and the significant potential for providing the most cost-efficient lipid therapy (Table 4), the formulary conversion from rosuvastatin to atorvastatin was warranted.

Limitations

There are several limitations for extrapolating results from this study to the general population. Due to the retrospective design of the study, no formal assessment of adherence was conducted. A prospective trial, with a researcher monitoring refills and tablet counts would be more accurate to ensure patients adherence with statin therapy.

Many of the patients that were included in the formulary conversion did not have follow-up laboratory work at the time of this study and were therefore excluded, leading to a smaller study population. A larger study population with a similar study design may be able to detect more significant correlations.

There are also several potential complications in regard to formulary conversions, including the inability to ensure the exact period that the patient switched therapies. During the conversion of rosuvastatin to atorvastatin at HVAMC, an attempt was made to minimize this confounder by converting patients on request for a refill of their rosuvastatin therapy. Last, all 185 patients that were studied were male; therefore, it is difficult to extrapolate these results for female patients.

Conclusions

This study shows that the conversion of patients from rosuvastatin therapy to atorvastatin was effective when it targeted a specific LDL-C goal or specific reduction in LDL-C. A similar conversion would likely lead to lower drug costs for many other health systems. Additional studies will be necessary given the recent changes in the national lipid guidelines. Furthermore, studies will be needed to assess concrete clinical endpoints (cardiovascular mortality, all-cause mortality, and cardiac events).

Related: Poor Outcomes for African Americans in Cardiac Rehabilitation

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Vancomycin was isolated in the 1950s, but due to impurities causing adverse events and semisynthetic penicillin production, its use was greatly reduced.^1,2 However, this medication gained in popularity 30 years later as a first-line treatment for methicillin-resistant Staphylococcus aureus infections.

In 2009 the Infectious Diseases Society of America (IDSA), American Society of Health System Pharmacists, and Society of Infectious Diseases Pharmacists developed a consensus review of the therapeutic monitoring and dosing of vancomycin in adult patients.³ Trough serum concentration levels are recommended as the most accurate and convenient method to monitor vancomycin. Per IDSA guidelines, an optimal trough is intended to be high enough to clear infections (> 10 mg/L) and prevent the development of vancomycin intermediate and resistant bacteria. Troughs should be obtained just before the next dose in steady-state conditions (starting just before the fourth dose) in patients with normal renal function.

Since the development of these guidelines, vancomycin trough levels are often drawn early.^4-7 This may lead to an overestimation of the true trough concentration. A study by Morrison and colleagues in Boston, Massachusetts, found that 41.3% of vancomycin troughs were drawn early, and this resulted in statistically significant increases in the vancomycin concentrations, the rate of vancomycin regimen adjustments (decrease, discontinuation, or holding of dose), and the repeat vancomycin level orders compared with correctly timed troughs.⁵ It was noted by the study authors that lowering the daily dose of vancomycin based on early trough levels could lead to an underdosing of vancomycin and an increase in intermediate or resistant bacteria.

Related: IDWEEK: Antibiotic ‘time-out’ cut vancomycin use 

The prevalence and implications of early trough samples have been measured at only 1 facility, and it is unknown whether these data can be reproduced elsewhere.⁵ Thus, this study sought to determine the prevalence and corresponding clinical actions of early trough levels at the Captain James A. Lovell Federal Health Care Center (JALFHCC). This is a unique facility that in 2010 combined a VA hospital with a DoD hospital. This facility cares for 67,000 military and retiree beneficiaries each year from southwestern Wisconsin and northwestern Illinois.The primary objective of this study was to measure the rate of early troughs drawn and their resultant effect on vancomycin regimens compared with correctly timed troughs. Secondarily, this study sought to compare the rate of repeated vancomycin trough levels in early vs correctly timed measurements.

Methods

This retrospective cohort analysis compared the outcomes of early and correctly timed vancomycin troughs. This study was approved by the Edward Hines, Jr. VA Hospital and JALFHCC Institutional Review Board. Veteran patients aged ≥ 18 years, hospitalized at JALFHCC, and receiving IV vancomycin at dosing intervals of 8, 12, 24, and 48 hours with measured trough levels between July 1, 2009, and July 1, 2013, were included in this study. Patients were excluded from analysis if vancomycin was given at any schedule other than the previously stated frequencies, they received hemodialysis during the treatment period, or their insurance coverage was through TRICARE (these patients had either active-duty or retired active-duty status).

Potentially eligible patients were identified via a Computerized Patient Records System (CPRS) search for laboratory vancomycin level measurements. The search supplied the researcher with the patient name, vancomycin level date and time, type of vancomycin level (trough or random), and vancomycin concentration. With this information, further data were gathered through CPRS: demographics, type of clinical infection, desired trough level (inferred if not listed in CPRS note), and vancomycin administration time (through the bar code medication administration system [BCMA] in CPRS). This analysis was of troughs, and multiple troughs may have originated from the same patient.

An early trough was defined as a trough taken more than 2 hours earlier than the next theoretical administration time or anytime before the third dose. After a trough was determined to be early or on time, the clinical actions taken during the dosing interval following sample collection were documented. A dose was considered to be held if stated in the BCMA or in a CPRS provider note. A dose was considered to be decreased with a change in frequency or strength that resulted in an overall daily dose decrease. A recollected vancomycin trough was counted within 24 hours of the trough or per a note in CPRS. Finally, observations that noted trends in vancomycin trough management were recorded.

The chi-square test with a significance criterion of 0.05 was used to compare early and on time troughs. Based on the results from the Boston, Massachusetts, study and 1 other study, about 780 vancomycin troughs would be required to meet significance in the primary outcome.^5,6

Results

A total of 474 patient charts were reviewed, and 278 met inclusion criteria (196 were excluded). Of the included patients, 799 trough levels were analyzed. Of these, 377 (42.2%) were drawn early. There was no significant difference in the baseline characteristics of the early group vs the correctly timed group (Table 1). Of the early troughs, 190 (56.3%) were drawn prior to the third dose of vancomycin. It was observed that a large portion of these troughs occurred after a vancomycin dose adjustment.

Clinical actions taken after sampling occurred at a rate of 14.5% in the early group and 22.9% in the correctly timed group (P = .003; Table 2). Early troughs led to a 7.7% rate of trough recollection, which was significantly greater than the 1.5% rate in the correctly timed group (P < .001). An analysis of each factor resulting in a clinical action illustrated that the rates of daily dose decrease and discontinued dose were similar between the groups (Table 3). However, the rate of held doses was 8.3% in the early group and 17.1% in the correctly timed group.

This research process yielded some observations. Occasionally a trough was drawn after vancomycin therapy was discontinued and when there was no concern for nephrotoxicity. After the guidelines were published, providers continued to document in CPRS notes to check troughs before the third dose. This incidence decreased over time. Troughs were taken often in patients who were receiving a short course of therapy or who were hemodynamically stable. Finally, documentation of vancomycin regimen changes occasionally did not match the record in the BCMA (in these situations, the BCMA record was used for this study).

Related: Assessment of High Staphylococcus aureus MIC and Poor Patient Outcomes

Discussion

A large portion of trough levels at the JALFHCC were drawn early and did not adhere to the 2009 consensus guidelines. The rates of early troughs in this study and in the Boston, Massachusetts, study are similar.⁵ However, the 2 studies differed in 1 significant aspect: Clinical actions were taken less often in the early group at JALFHCC, whereas they were taken more often in the early group in the Boston, Massachusetts, study. This dissimilarity could be attributed to a difference in software between the hospitals. In the previous study, trough levels and the time that they were drawn were not displayed together. Thus, clinicians may have been less likely to gauge whether a trough was early. Since this information is available at the JALFHCC, clinicians may have been aware that the trough was early and avoided adjusting treatment (such as holding a dose, as illustrated in the data) based on a falsely elevated trough. This point is further supported by significantly greater amounts of recollected troughs in the early group, suggesting an understanding that the trough was early.

The low trough recollection rate of 7.7% of all early samples could be due to several factors that would prevent a trough redraw. First, medication discontinuation resulting from course completion or sensitivity results would not require further trough monitoring. Second, practitioners may assess the early sample as insignificantly different from a correctly timed one and elect not to redraw the trough. Sometimes a trough was drawn at the correct time, but the time was recorded incorrectly. In this situation, a new trough level would not be necessary. Finally, a lack of sufficient staffing during nights and weekends may result in a delay in interpreting results leading to a missed opportunity for recollection. Additionally, some troughs may not have been redrawn based on a practitioner’s opinion that a trough was not significantly early and did not represent skewed results. Sometimes an incorrect recording of trough draw time reflected that it was taken after vancomycin dosing when it was not.

Specific observations regarding the timing of the trough indicate other possible concerns and areas for improvement. First, providers must cancel future trough orders concurrently with canceling treatment. Second, at the time of publication of the consensus, some providers were slow adopters of the new guidelines. Finally, the IDSA guidelines state that frequent monitoring for short course, lower intensity therapy, or in patients who are hemodynamically stable is not recommended.³ However, troughs were sometimes measured 2 to 3 times weekly in these patients.

Related: Results mixed in hospital efforts to tackle antimicrobial resistance

The data and observations lead to the conclusion that although providers may be able to discern between early and correctly timed troughs, they were not consistently adherent to the 2009 IDSA guidelines. It has been shown that pharmacy involvement of Medicare patients with infections in the intensive care unit has led to better clinical and monetary outcomes.⁸ Therefore, continued efforts by clinical pharmacists to monitor trough timing can be used to improve adherence and decrease costs (each trough is estimated to cost $16.97).

A study conducted in Australia demonstrated that pharmacist-led education of vancomycin dosing and monitoring (including when to measure a trough level) among prescribers and nurses led to improved adherence to the current guidelines and a greater number of patients treated within desired therapeutic ranges.⁹ In addition, a small study at the Atlanta VAMC in Georgia demonstrated that education of nurses, lab personnel, residents, ward clerks, and pharmacists led to a greater number of appropriately timed vancomycin and aminoglycoside levels.¹⁰ Thus, an interdisciplinary review of the current IDSA guidelines and review on the publication of the anticipated updated vancomycin guidelines should be provided to hospital personnel to aid in adoption of current dosing and monitoring recommendations.¹¹

Limitations

This study is limited by the 4-year span of time that it encompassed, which may give a skewed depiction of current practices. Another limitation is that patients with fluctuating renal function were included in the analysis. Instead of selecting a random level order, a trough level order was sometimes selected for these patients. This could lead to a lower actual rate of early troughs. A third limitation is that this was a small and unblinded study. Also, the actual trough levels and the resulting changes that were made to specific regimens were not recorded. Thus, these data do not indicate whether the changes that were made reflected guideline recommendations. Finally, some clinical actions were taken after the dosing interval following the trough. This was often a result of off-hours lab results or waiting on attending physician or infectious disease guidance. These data were not included in the analysis.

Conclusion

Vancomycin troughs were often drawn too early and resulted in an increased rate of trough recollection. In an attempt to improve adherence to the current and the upcoming revised version of the IDSA consensus statement, it is recommended to educate and reeducate providers through interdisciplinary-led review sessions.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Vancomycin was isolated in the 1950s, but due to impurities causing adverse events and semisynthetic penicillin production, its use was greatly reduced.^1,2 However, this medication gained in popularity 30 years later as a first-line treatment for methicillin-resistant Staphylococcus aureus infections.

In 2009 the Infectious Diseases Society of America (IDSA), American Society of Health System Pharmacists, and Society of Infectious Diseases Pharmacists developed a consensus review of the therapeutic monitoring and dosing of vancomycin in adult patients.³ Trough serum concentration levels are recommended as the most accurate and convenient method to monitor vancomycin. Per IDSA guidelines, an optimal trough is intended to be high enough to clear infections (> 10 mg/L) and prevent the development of vancomycin intermediate and resistant bacteria. Troughs should be obtained just before the next dose in steady-state conditions (starting just before the fourth dose) in patients with normal renal function.

Since the development of these guidelines, vancomycin trough levels are often drawn early.^4-7 This may lead to an overestimation of the true trough concentration. A study by Morrison and colleagues in Boston, Massachusetts, found that 41.3% of vancomycin troughs were drawn early, and this resulted in statistically significant increases in the vancomycin concentrations, the rate of vancomycin regimen adjustments (decrease, discontinuation, or holding of dose), and the repeat vancomycin level orders compared with correctly timed troughs.⁵ It was noted by the study authors that lowering the daily dose of vancomycin based on early trough levels could lead to an underdosing of vancomycin and an increase in intermediate or resistant bacteria.

Related: IDWEEK: Antibiotic ‘time-out’ cut vancomycin use 

The prevalence and implications of early trough samples have been measured at only 1 facility, and it is unknown whether these data can be reproduced elsewhere.⁵ Thus, this study sought to determine the prevalence and corresponding clinical actions of early trough levels at the Captain James A. Lovell Federal Health Care Center (JALFHCC). This is a unique facility that in 2010 combined a VA hospital with a DoD hospital. This facility cares for 67,000 military and retiree beneficiaries each year from southwestern Wisconsin and northwestern Illinois.The primary objective of this study was to measure the rate of early troughs drawn and their resultant effect on vancomycin regimens compared with correctly timed troughs. Secondarily, this study sought to compare the rate of repeated vancomycin trough levels in early vs correctly timed measurements.

Methods

This retrospective cohort analysis compared the outcomes of early and correctly timed vancomycin troughs. This study was approved by the Edward Hines, Jr. VA Hospital and JALFHCC Institutional Review Board. Veteran patients aged ≥ 18 years, hospitalized at JALFHCC, and receiving IV vancomycin at dosing intervals of 8, 12, 24, and 48 hours with measured trough levels between July 1, 2009, and July 1, 2013, were included in this study. Patients were excluded from analysis if vancomycin was given at any schedule other than the previously stated frequencies, they received hemodialysis during the treatment period, or their insurance coverage was through TRICARE (these patients had either active-duty or retired active-duty status).

Potentially eligible patients were identified via a Computerized Patient Records System (CPRS) search for laboratory vancomycin level measurements. The search supplied the researcher with the patient name, vancomycin level date and time, type of vancomycin level (trough or random), and vancomycin concentration. With this information, further data were gathered through CPRS: demographics, type of clinical infection, desired trough level (inferred if not listed in CPRS note), and vancomycin administration time (through the bar code medication administration system [BCMA] in CPRS). This analysis was of troughs, and multiple troughs may have originated from the same patient.

An early trough was defined as a trough taken more than 2 hours earlier than the next theoretical administration time or anytime before the third dose. After a trough was determined to be early or on time, the clinical actions taken during the dosing interval following sample collection were documented. A dose was considered to be held if stated in the BCMA or in a CPRS provider note. A dose was considered to be decreased with a change in frequency or strength that resulted in an overall daily dose decrease. A recollected vancomycin trough was counted within 24 hours of the trough or per a note in CPRS. Finally, observations that noted trends in vancomycin trough management were recorded.

The chi-square test with a significance criterion of 0.05 was used to compare early and on time troughs. Based on the results from the Boston, Massachusetts, study and 1 other study, about 780 vancomycin troughs would be required to meet significance in the primary outcome.^5,6

Results

A total of 474 patient charts were reviewed, and 278 met inclusion criteria (196 were excluded). Of the included patients, 799 trough levels were analyzed. Of these, 377 (42.2%) were drawn early. There was no significant difference in the baseline characteristics of the early group vs the correctly timed group (Table 1). Of the early troughs, 190 (56.3%) were drawn prior to the third dose of vancomycin. It was observed that a large portion of these troughs occurred after a vancomycin dose adjustment.

Clinical actions taken after sampling occurred at a rate of 14.5% in the early group and 22.9% in the correctly timed group (P = .003; Table 2). Early troughs led to a 7.7% rate of trough recollection, which was significantly greater than the 1.5% rate in the correctly timed group (P < .001). An analysis of each factor resulting in a clinical action illustrated that the rates of daily dose decrease and discontinued dose were similar between the groups (Table 3). However, the rate of held doses was 8.3% in the early group and 17.1% in the correctly timed group.

This research process yielded some observations. Occasionally a trough was drawn after vancomycin therapy was discontinued and when there was no concern for nephrotoxicity. After the guidelines were published, providers continued to document in CPRS notes to check troughs before the third dose. This incidence decreased over time. Troughs were taken often in patients who were receiving a short course of therapy or who were hemodynamically stable. Finally, documentation of vancomycin regimen changes occasionally did not match the record in the BCMA (in these situations, the BCMA record was used for this study).

Related: Assessment of High Staphylococcus aureus MIC and Poor Patient Outcomes

Discussion

A large portion of trough levels at the JALFHCC were drawn early and did not adhere to the 2009 consensus guidelines. The rates of early troughs in this study and in the Boston, Massachusetts, study are similar.⁵ However, the 2 studies differed in 1 significant aspect: Clinical actions were taken less often in the early group at JALFHCC, whereas they were taken more often in the early group in the Boston, Massachusetts, study. This dissimilarity could be attributed to a difference in software between the hospitals. In the previous study, trough levels and the time that they were drawn were not displayed together. Thus, clinicians may have been less likely to gauge whether a trough was early. Since this information is available at the JALFHCC, clinicians may have been aware that the trough was early and avoided adjusting treatment (such as holding a dose, as illustrated in the data) based on a falsely elevated trough. This point is further supported by significantly greater amounts of recollected troughs in the early group, suggesting an understanding that the trough was early.

The low trough recollection rate of 7.7% of all early samples could be due to several factors that would prevent a trough redraw. First, medication discontinuation resulting from course completion or sensitivity results would not require further trough monitoring. Second, practitioners may assess the early sample as insignificantly different from a correctly timed one and elect not to redraw the trough. Sometimes a trough was drawn at the correct time, but the time was recorded incorrectly. In this situation, a new trough level would not be necessary. Finally, a lack of sufficient staffing during nights and weekends may result in a delay in interpreting results leading to a missed opportunity for recollection. Additionally, some troughs may not have been redrawn based on a practitioner’s opinion that a trough was not significantly early and did not represent skewed results. Sometimes an incorrect recording of trough draw time reflected that it was taken after vancomycin dosing when it was not.

Specific observations regarding the timing of the trough indicate other possible concerns and areas for improvement. First, providers must cancel future trough orders concurrently with canceling treatment. Second, at the time of publication of the consensus, some providers were slow adopters of the new guidelines. Finally, the IDSA guidelines state that frequent monitoring for short course, lower intensity therapy, or in patients who are hemodynamically stable is not recommended.³ However, troughs were sometimes measured 2 to 3 times weekly in these patients.

Related: Results mixed in hospital efforts to tackle antimicrobial resistance

The data and observations lead to the conclusion that although providers may be able to discern between early and correctly timed troughs, they were not consistently adherent to the 2009 IDSA guidelines. It has been shown that pharmacy involvement of Medicare patients with infections in the intensive care unit has led to better clinical and monetary outcomes.⁸ Therefore, continued efforts by clinical pharmacists to monitor trough timing can be used to improve adherence and decrease costs (each trough is estimated to cost $16.97).

A study conducted in Australia demonstrated that pharmacist-led education of vancomycin dosing and monitoring (including when to measure a trough level) among prescribers and nurses led to improved adherence to the current guidelines and a greater number of patients treated within desired therapeutic ranges.⁹ In addition, a small study at the Atlanta VAMC in Georgia demonstrated that education of nurses, lab personnel, residents, ward clerks, and pharmacists led to a greater number of appropriately timed vancomycin and aminoglycoside levels.¹⁰ Thus, an interdisciplinary review of the current IDSA guidelines and review on the publication of the anticipated updated vancomycin guidelines should be provided to hospital personnel to aid in adoption of current dosing and monitoring recommendations.¹¹

Limitations

This study is limited by the 4-year span of time that it encompassed, which may give a skewed depiction of current practices. Another limitation is that patients with fluctuating renal function were included in the analysis. Instead of selecting a random level order, a trough level order was sometimes selected for these patients. This could lead to a lower actual rate of early troughs. A third limitation is that this was a small and unblinded study. Also, the actual trough levels and the resulting changes that were made to specific regimens were not recorded. Thus, these data do not indicate whether the changes that were made reflected guideline recommendations. Finally, some clinical actions were taken after the dosing interval following the trough. This was often a result of off-hours lab results or waiting on attending physician or infectious disease guidance. These data were not included in the analysis.

Conclusion

Vancomycin troughs were often drawn too early and resulted in an increased rate of trough recollection. In an attempt to improve adherence to the current and the upcoming revised version of the IDSA consensus statement, it is recommended to educate and reeducate providers through interdisciplinary-led review sessions.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Vancomycin was isolated in the 1950s, but due to impurities causing adverse events and semisynthetic penicillin production, its use was greatly reduced.^1,2 However, this medication gained in popularity 30 years later as a first-line treatment for methicillin-resistant Staphylococcus aureus infections.

In 2009 the Infectious Diseases Society of America (IDSA), American Society of Health System Pharmacists, and Society of Infectious Diseases Pharmacists developed a consensus review of the therapeutic monitoring and dosing of vancomycin in adult patients.³ Trough serum concentration levels are recommended as the most accurate and convenient method to monitor vancomycin. Per IDSA guidelines, an optimal trough is intended to be high enough to clear infections (> 10 mg/L) and prevent the development of vancomycin intermediate and resistant bacteria. Troughs should be obtained just before the next dose in steady-state conditions (starting just before the fourth dose) in patients with normal renal function.

Since the development of these guidelines, vancomycin trough levels are often drawn early.^4-7 This may lead to an overestimation of the true trough concentration. A study by Morrison and colleagues in Boston, Massachusetts, found that 41.3% of vancomycin troughs were drawn early, and this resulted in statistically significant increases in the vancomycin concentrations, the rate of vancomycin regimen adjustments (decrease, discontinuation, or holding of dose), and the repeat vancomycin level orders compared with correctly timed troughs.⁵ It was noted by the study authors that lowering the daily dose of vancomycin based on early trough levels could lead to an underdosing of vancomycin and an increase in intermediate or resistant bacteria.

Related: IDWEEK: Antibiotic ‘time-out’ cut vancomycin use 

The prevalence and implications of early trough samples have been measured at only 1 facility, and it is unknown whether these data can be reproduced elsewhere.⁵ Thus, this study sought to determine the prevalence and corresponding clinical actions of early trough levels at the Captain James A. Lovell Federal Health Care Center (JALFHCC). This is a unique facility that in 2010 combined a VA hospital with a DoD hospital. This facility cares for 67,000 military and retiree beneficiaries each year from southwestern Wisconsin and northwestern Illinois.The primary objective of this study was to measure the rate of early troughs drawn and their resultant effect on vancomycin regimens compared with correctly timed troughs. Secondarily, this study sought to compare the rate of repeated vancomycin trough levels in early vs correctly timed measurements.

Methods

This retrospective cohort analysis compared the outcomes of early and correctly timed vancomycin troughs. This study was approved by the Edward Hines, Jr. VA Hospital and JALFHCC Institutional Review Board. Veteran patients aged ≥ 18 years, hospitalized at JALFHCC, and receiving IV vancomycin at dosing intervals of 8, 12, 24, and 48 hours with measured trough levels between July 1, 2009, and July 1, 2013, were included in this study. Patients were excluded from analysis if vancomycin was given at any schedule other than the previously stated frequencies, they received hemodialysis during the treatment period, or their insurance coverage was through TRICARE (these patients had either active-duty or retired active-duty status).

Potentially eligible patients were identified via a Computerized Patient Records System (CPRS) search for laboratory vancomycin level measurements. The search supplied the researcher with the patient name, vancomycin level date and time, type of vancomycin level (trough or random), and vancomycin concentration. With this information, further data were gathered through CPRS: demographics, type of clinical infection, desired trough level (inferred if not listed in CPRS note), and vancomycin administration time (through the bar code medication administration system [BCMA] in CPRS). This analysis was of troughs, and multiple troughs may have originated from the same patient.

An early trough was defined as a trough taken more than 2 hours earlier than the next theoretical administration time or anytime before the third dose. After a trough was determined to be early or on time, the clinical actions taken during the dosing interval following sample collection were documented. A dose was considered to be held if stated in the BCMA or in a CPRS provider note. A dose was considered to be decreased with a change in frequency or strength that resulted in an overall daily dose decrease. A recollected vancomycin trough was counted within 24 hours of the trough or per a note in CPRS. Finally, observations that noted trends in vancomycin trough management were recorded.

The chi-square test with a significance criterion of 0.05 was used to compare early and on time troughs. Based on the results from the Boston, Massachusetts, study and 1 other study, about 780 vancomycin troughs would be required to meet significance in the primary outcome.^5,6

Results

A total of 474 patient charts were reviewed, and 278 met inclusion criteria (196 were excluded). Of the included patients, 799 trough levels were analyzed. Of these, 377 (42.2%) were drawn early. There was no significant difference in the baseline characteristics of the early group vs the correctly timed group (Table 1). Of the early troughs, 190 (56.3%) were drawn prior to the third dose of vancomycin. It was observed that a large portion of these troughs occurred after a vancomycin dose adjustment.

Clinical actions taken after sampling occurred at a rate of 14.5% in the early group and 22.9% in the correctly timed group (P = .003; Table 2). Early troughs led to a 7.7% rate of trough recollection, which was significantly greater than the 1.5% rate in the correctly timed group (P < .001). An analysis of each factor resulting in a clinical action illustrated that the rates of daily dose decrease and discontinued dose were similar between the groups (Table 3). However, the rate of held doses was 8.3% in the early group and 17.1% in the correctly timed group.

This research process yielded some observations. Occasionally a trough was drawn after vancomycin therapy was discontinued and when there was no concern for nephrotoxicity. After the guidelines were published, providers continued to document in CPRS notes to check troughs before the third dose. This incidence decreased over time. Troughs were taken often in patients who were receiving a short course of therapy or who were hemodynamically stable. Finally, documentation of vancomycin regimen changes occasionally did not match the record in the BCMA (in these situations, the BCMA record was used for this study).

Related: Assessment of High Staphylococcus aureus MIC and Poor Patient Outcomes

Discussion

A large portion of trough levels at the JALFHCC were drawn early and did not adhere to the 2009 consensus guidelines. The rates of early troughs in this study and in the Boston, Massachusetts, study are similar.⁵ However, the 2 studies differed in 1 significant aspect: Clinical actions were taken less often in the early group at JALFHCC, whereas they were taken more often in the early group in the Boston, Massachusetts, study. This dissimilarity could be attributed to a difference in software between the hospitals. In the previous study, trough levels and the time that they were drawn were not displayed together. Thus, clinicians may have been less likely to gauge whether a trough was early. Since this information is available at the JALFHCC, clinicians may have been aware that the trough was early and avoided adjusting treatment (such as holding a dose, as illustrated in the data) based on a falsely elevated trough. This point is further supported by significantly greater amounts of recollected troughs in the early group, suggesting an understanding that the trough was early.

The low trough recollection rate of 7.7% of all early samples could be due to several factors that would prevent a trough redraw. First, medication discontinuation resulting from course completion or sensitivity results would not require further trough monitoring. Second, practitioners may assess the early sample as insignificantly different from a correctly timed one and elect not to redraw the trough. Sometimes a trough was drawn at the correct time, but the time was recorded incorrectly. In this situation, a new trough level would not be necessary. Finally, a lack of sufficient staffing during nights and weekends may result in a delay in interpreting results leading to a missed opportunity for recollection. Additionally, some troughs may not have been redrawn based on a practitioner’s opinion that a trough was not significantly early and did not represent skewed results. Sometimes an incorrect recording of trough draw time reflected that it was taken after vancomycin dosing when it was not.

Specific observations regarding the timing of the trough indicate other possible concerns and areas for improvement. First, providers must cancel future trough orders concurrently with canceling treatment. Second, at the time of publication of the consensus, some providers were slow adopters of the new guidelines. Finally, the IDSA guidelines state that frequent monitoring for short course, lower intensity therapy, or in patients who are hemodynamically stable is not recommended.³ However, troughs were sometimes measured 2 to 3 times weekly in these patients.

Related: Results mixed in hospital efforts to tackle antimicrobial resistance

The data and observations lead to the conclusion that although providers may be able to discern between early and correctly timed troughs, they were not consistently adherent to the 2009 IDSA guidelines. It has been shown that pharmacy involvement of Medicare patients with infections in the intensive care unit has led to better clinical and monetary outcomes.⁸ Therefore, continued efforts by clinical pharmacists to monitor trough timing can be used to improve adherence and decrease costs (each trough is estimated to cost $16.97).

A study conducted in Australia demonstrated that pharmacist-led education of vancomycin dosing and monitoring (including when to measure a trough level) among prescribers and nurses led to improved adherence to the current guidelines and a greater number of patients treated within desired therapeutic ranges.⁹ In addition, a small study at the Atlanta VAMC in Georgia demonstrated that education of nurses, lab personnel, residents, ward clerks, and pharmacists led to a greater number of appropriately timed vancomycin and aminoglycoside levels.¹⁰ Thus, an interdisciplinary review of the current IDSA guidelines and review on the publication of the anticipated updated vancomycin guidelines should be provided to hospital personnel to aid in adoption of current dosing and monitoring recommendations.¹¹

Limitations

This study is limited by the 4-year span of time that it encompassed, which may give a skewed depiction of current practices. Another limitation is that patients with fluctuating renal function were included in the analysis. Instead of selecting a random level order, a trough level order was sometimes selected for these patients. This could lead to a lower actual rate of early troughs. A third limitation is that this was a small and unblinded study. Also, the actual trough levels and the resulting changes that were made to specific regimens were not recorded. Thus, these data do not indicate whether the changes that were made reflected guideline recommendations. Finally, some clinical actions were taken after the dosing interval following the trough. This was often a result of off-hours lab results or waiting on attending physician or infectious disease guidance. These data were not included in the analysis.

Conclusion

Vancomycin troughs were often drawn too early and resulted in an increased rate of trough recollection. In an attempt to improve adherence to the current and the upcoming revised version of the IDSA consensus statement, it is recommended to educate and reeducate providers through interdisciplinary-led review sessions.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

In 2012, 9.3% of the U.S. population had diabetes mellitus (DM).¹ According to the American Diabetes Association, in 2012, the total cost of diagnosed DM in the U.S. was $245 billion.² Diabetes mellitus is a leading cause of blindness, end-stage renal disease, and amputation in the U.S.³ Up to 80% of patients with DM will develop or die of macrovascular disease, such as heart attack or stroke.³

Diabetes mellitus is a chronic disease of epidemic proportion with management complexity that threatens to overwhelm providers in the acute care and primary care settings. Limited specialist availability and increased wait times continue to afflict the VA health care system, prompting efforts to increase health care provider (HCP) access and improve clinic efficiency.⁴ One of the methods proposed to increase HCP access and maximize clinic efficiency is the shared medical appointment (SMA).^5,6

The SMA was designed to improve access and quality of care through enhanced education and support. With the number of people living with chronic diseases on the rise, the current patient-provider model is unrealistic in today’s health care environment. Shared medical appointments offer a unique format for providing evidence-based chronic disease management in which patients and a multidisciplinary team of providers collaborate toward education, discussion, and medication management in a supportive environment.⁷ Research has suggested that SMAs are a successful way to manage type 2 DM (T2DM).^8,9 However, there is uncertainty regarding the optimal model design. The goals of this study were to evaluate whether the diabetes SMA at the Adam Benjamin, Jr. (ABJ) community-based outpatient clinic (CBOC) was an effective practice model for achieving improvements in glycemic control and to use subgroup analyses to elucidate unique characteristics about SMAs that may have been correlated with clinical success. This study may provide valuable information for other facilities considering SMAs.

Overview

The Jesse Brown VAMC (JBVAMC) and the ABJ CBOC implemented a T2DM-focused SMA in 2011. The ABJ CBOC multidisciplinary SMA team consisted of a medical administration service clerk, a registered dietician, a certified DM educator, a registered nurse, a nurse practitioner (NP), and a clinical pharmacy specialist (CPS). This team collaborated to deliver high-quality care to patients with poorly controlled T2DM to improve their glycemic control as well as clinical knowledge of their disease. A private conference room at the ABJ CBOC served as the location for the SMAs. This room was divided into 2 adjacent areas: One area with tables was organized in a semicircle to promote group discussion as well as minimize isolated conversations; the other area had computer terminals to facilitate individualized medication management. Other equipment included a scale for obtaining patient weights and various audio-visual devices.

Related: Efficacy of Patient Aligned Care Team Pharmacist Services in Reaching Diabetes and Hyperlipidemia Treatment Goals

The ABJ CBOC offered monthly SMAs. The team made several attempts to maximize SMA show rates, as previous studies indicated that low SMA show rates were a barrier to success.^3,4,7-9 One review reported no-show rates as high as 70% in certain group visit models.⁴ About 2 weeks prior to a session, prospective SMA patients received automated and customized preappointment letters. Automated and customized phone call reminders were made to prospective SMA patients a few days before each session. As many as 18 patients participated in a single ABJ SMA.

The ABJ SMAs lasted from 60 to 90 minutes, depending on the level of patient participation and the size of the group. The first half of the SMA was dedicated to a group discussion, which involved the SMA team, the patient, and the patient’s family (if desired). The topic of conversation was typically guided by patient curiosity and knowledge deficits in a spontaneous and free-flowing manner; for this reason, these sessions were considered to be open.

The team also engaged in more structured focused sessions, which limited the spontaneous flow of conservation and narrowed the scope to provide targeted education about various aspects of T2DM care. During focused sessions, services such as dental, optometry, podiatry, MOVE! (a VA self-management weight reduction program), and nutrition also participated. Focused sessions addressed topics such as hypoglycemia management, eating around the holidays, sick-day management of T2DM, grocery shopping, exercise, oral health, eye care, and foot care. The specialty services were encouraged to be creative and interactive during the SMA. Many of these services used supportive literature, demonstrations, diagrams, and props to enrich the educational experience. Group discussion typically lasted 30 to 40 minutes; after which patients met individually with either a CPS or NP for medication management.

Medication management focused on optimizing T2DM therapy (both oral and injectable) to improve glycemic control. Interventions outside of T2DM therapy (eg, cholesterol, hypertension, and other risk reduction modalities) were not made, due to time constraints. Once a patient demonstrated improved working knowledge of T2DM and a clinically significant reduction in their glycosylated hemoglobin A_1c (A_1c) they were discharged from SMAs at the discretion of the SMA team. There was no set minimum or maximum duration for the SMAs.

Methods

This study was a retrospective chart review conducted at the JBVAMC and was approved by the institutional review board and the research and development committee. Patient confidentiality was maintained by identifying patients by means other than name or unique identifiers. Protected health information was accessible only by the aforementioned investigators. There was no direct patient contact during this study.

Patient lists were generated from the computerized patient record system (CPRS). Patients were tracked up to 6 months after SMA discharge or until the last SMA in which they participated. The control group was matched according to location, age, glycemic control, and time. The control group never attended an ABJ SMA but may have received regular care through their primary care provider, CPS, or endocrinologist. Prospective control group patients were randomized and reviewed sequentially to obtain the matched cohort.

The study took place at ABJ, an outpatient clinic serving veterans in northwest Indiana and surrounding areas. Inclusion criteria for the SMA group were patients with T2DM, aged ≥ 45 years, with an A_1c ≥ 8.5% seen at ABJ for T2DM from May 1, 2011, to June 30, 2013. The control group included patients with T2DM, aged ≥ 45 years, with an A_1c > 9% who never attended SMAs but may have received regular care at ABJ during the study period. The SMA group’s inclusion criteria threshold for A_1c was lower in order to maximize sample size. The control group’s inclusion criteria threshold for A_1c was higher due to use of a default reminder report called “A_1c > 9%” to generate patient lists. Patients were excluded from the study if they did not meet inclusion criteria.

Baseline datum was the most recent parameter available in CPRS prior to enrollment. The endpoint datum was the parameter nearest the time of SMA discharge or the first available parameter within 6 months from the date of discharge. In the control group, the baseline datum was the initial parameter during the study period and the endpoint datum was the closest measure to 4 months after baseline. Four months was chosen to allow for at least 1 A_1c measurement during the study period. In addition, it was estimated (prior to collecting any data) that 4 months was the average time a patient participated in SMAs. Serial A_1c measurements were defined as values obtained at SMA discharge and 3- and 6-months postdischarge. These parameters were used to evaluate the sustainability of improvements in glycemic control. All values falling outside of these defined parameters were excluded.

Related: Experiences of Veterans With Diabetes From Shared Medical Appointments

The data analysis compared A_1c change from baseline to endpoint for the SMA and control groups. Data collection included baseline characteristics, SMA show rate, number of SMA patients seen by a CPS or NP, number and type of SMA interventions made by a CPS or NP, and the number and type of non-SMA interventions made during the study period. Intervention types were medications: added, discontinued, or titrated; and other, defined as referrals made to specialty services (such as dental, optometry, and podiatry).

Secondary endpoints included the number of SMAs and glycemic improvement, SMA format style (open- vs focused session) and glycemic improvement, SMA provider (CPS vs NP) and glycemic improvement, the change in A_1c stratified by baseline A_1c (A_1c ≥ 10% vs < 10%), the change in actual body weight (ABW) and body mass index (BMI), and maintenance of A_1c (3- and 6-months postdischarge).

The primary endpoint was evaluated using a 2-sample Student t test. Secondary endpoints were evaluated using the independent t test. Statistical significance was defined as P < .05.

Results

A total of 129 unique patients were scheduled for SMAs, 62 of which met inclusion criteria and were included in the SMA group. During enrollment, 67 patients were excluded: 55 never participated in SMAs, 6 had baseline A_1c values < 8.5%, 4 had insufficient data, and 2 were aged < 45 years. A total of 29 SMAs were conducted during the study period, and patients attended an average of 3.15 ± 2.14 (SD) SMAs. The average attendance at each SMA was 7.1 ± 2.62 (SD) patients. For the control group, 754 unique patients were identified and randomized. A total of 90 charts were sequentially reviewed in order to obtain the 62 patients for the control group.

Baseline characteristics were balanced between groups. However, there were more women in the SMA group vs the control group (Table 1). Within the control group, there were a total of 107 appointments that addressed T2DM, which averaged 1.72 ± 1.51 (SD) appointments per patient. The total number of interventions made in the SMA group was 192: 64.6% (124) by a CPS and 35.4% (68) by a NP. For the CPS, the most frequent intervention was medication titration (69.5%), followed by other (23.5%), medication addition (4%), and medication discontinuation (3%). Of note, 53.2% (33) of the SMA patients were seen an average of 1.2 times by non-SMA providers. The SMA patients had a total of 45 non-SMA interventions (0.73 per patient) during the study period.

For the primary endpoint, the SMA group had a 1.48% ± 0.02 (SD) reduction in A_1c compared with a 0.6% ± 0.02 (SD) decrease in the control group (P = .01). When evaluating mean changes in A_1c by the number of SMAs attended, it was noted that participation in ≥ 6 SMAs led to the greatest reduction in A_1c of 2.08%. In the SMA group, it was noted that patients with higher A_1c values at baseline demonstrated greater improvements in glycemic control compared with patients with lower baseline A_1c values. The mean change in A_1c, stratified by baseline A_1c, was -2.26% for those with baseline A_1c values ≥ 10% and -0.87% for those with baseline A_1c values < 10%.

In evaluating the format style, open- vs focused-session, it was observed that participation in focused sessions led to greater improvements in glycemic control. Furthermore, when stratified by provider, greater improvements in glycemic control were demonstrated when medication management was completed by a CPS vs a NP (Table 2). The average number of interventions per SMA patient was 3.1 ± 2.22 (SD). For the control group, the total number of interventions made was 86, with an average of 1.37 ± 1.51 (SD) per patient. The overall show rate was 49% ± 16 (SD), 52% ± 16 (SD) for open visits, and 46% ± 15 (SD) for focused visits. The mean change in ABW and BMI from baseline to endpoint was no different between the SMA and control groups (Table 3). The SMA group participants demonstrated a decrease in A_1c at 3 months postdischarge, and a moderate increase in A_1c was noted at 6 months postdischarge.

Discussion

Shared medical appointments provide an effective alternative to standards of care in order to obtain improvements in glycemic control. Consistent with previous studies, this study reported significant improvements in glycemic control in the SMA group vs the control group. This study also elucidated unique characteristics about SMAs that may have been correlated with clinical success.

Although the greatest improvements in glycemic control were noted for those who participated in ≥ 6 SMAs, it was observed that participation in only 1 SMA also led to improvements. For a site with limited staff and a high volume of patients waiting to participate in SMAs, it may be mutually beneficial to offer only 1 SMA per patient. In addition, patients with ≥ 10% A_1c at baseline demonstrated greater improvements in glycemic control compared to those with < 10% A_1c at baseline. The reasons the higher baseline A_1c subgroup responded to interventions more robustly are unclear and likely multifactorial. Nonetheless, factors such as psychosocial influences (eg, peer pressure to get healthy) may have increased motivation to prevent complications and improved medication adherence in the setting of closer follow-up. Additionally, hyperresponsiveness to drug therapy may have played a role. Regardless, for new SMA programs interested in making an immediate impact, it may be advantageous to initially select patients with very poorly controlled DM.

A unique aspect of the ABJ SMA was the variety of focused sessions offered. Previous studies did not demonstrate such a variety of focused sessions, nor did they evaluate the impact of a focused visit on the patient’s T2DM control. Participation in focused ABJ SMA sessions may have led to improved T2DM control, which may be attributed to the value patients assigned to specialty care and an increased motivation to get healthy.

Related: SGLT2 Inhibitors for Type 2 Diabetes Mellitus Treatment

Another factor that may have led to improved T2DM control was CPS involvement with medication management. The presence of a NP was highly valued, both from a group discussion and medication management standpoint; still, it is a good idea to involve a CPS who has a strong command of DM pharmacotherapy. One shortcoming of this SMA program was the inability for patients to maintain glycemic improvements 6 months after discharge. This pitfall was likely the result of suboptimal coordination of care after SMA discharge and may be avoided by asking the medical administration service clerk to promptly schedule discharged SMA patients for a general medicine clinic T2DM follow-up.The SMA patients had more T2DM interventions within the same time frame compared with the control patients. Although not causative, the increased number of interventions in addition to the bolstered support of the SMA may have correlated with glycemic improvements.

An important finding of this study was the SMA show rate and how it compared with attendance rates found in other group models. The favorable ABJ SMA show rate could have been due to the rigorous attention paid to reminder letters and phone calls. The literature has not established a standard approach to increasing SMA show rates; however, the current data suggest that increased reminders may have increased attendance.

Limitations

This study had several limitations. The external validity was weakened by the modest sample size and the homogenous baseline characteristics of those enrolled. Another limitation was inconsistent documentation of laboratory parameters. The inability to obtain A_1c values exactly at enrollment and discharge could have potentially skewed the results. In addition, incomplete documentation of interventions for dual-care patients (ie, those who obtained care outside of the VA) was an unavoidable challenge. Last, this study did not perform an assessment of SMA patient satisfaction, cost-benefit, or safety.

Conclusion

The ABJ SMA was an effective addition to standards of care in order to achieve improvements in glycemic control in a veteran population with poorly controlled T2DM. Furthermore,the data suggest that a successful program should be multidisciplinary, select poorly controlled patients, offer focused sessions, have a CPS participate in medication management, and encourage patients to complete ≥ 6 sessions. Future studies should be conducted to include more diverse patients to see whether the efficacy of this SMA format is maintained.

A safety analysis should also be conducted to ensure that the SMA format is not only effective, but also a safe means to manage medical conditions. In addition, the scope of the ABJ SMAs should be expanded to allow for evaluation of other diseases. An evaluation of patient satisfaction and cost-benefit could provide additional support for the implementation of SMAs, as improvements in quality of life and cost savings are endpoints to be desired.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

In 2012, 9.3% of the U.S. population had diabetes mellitus (DM).¹ According to the American Diabetes Association, in 2012, the total cost of diagnosed DM in the U.S. was $245 billion.² Diabetes mellitus is a leading cause of blindness, end-stage renal disease, and amputation in the U.S.³ Up to 80% of patients with DM will develop or die of macrovascular disease, such as heart attack or stroke.³

Diabetes mellitus is a chronic disease of epidemic proportion with management complexity that threatens to overwhelm providers in the acute care and primary care settings. Limited specialist availability and increased wait times continue to afflict the VA health care system, prompting efforts to increase health care provider (HCP) access and improve clinic efficiency.⁴ One of the methods proposed to increase HCP access and maximize clinic efficiency is the shared medical appointment (SMA).^5,6

The SMA was designed to improve access and quality of care through enhanced education and support. With the number of people living with chronic diseases on the rise, the current patient-provider model is unrealistic in today’s health care environment. Shared medical appointments offer a unique format for providing evidence-based chronic disease management in which patients and a multidisciplinary team of providers collaborate toward education, discussion, and medication management in a supportive environment.⁷ Research has suggested that SMAs are a successful way to manage type 2 DM (T2DM).^8,9 However, there is uncertainty regarding the optimal model design. The goals of this study were to evaluate whether the diabetes SMA at the Adam Benjamin, Jr. (ABJ) community-based outpatient clinic (CBOC) was an effective practice model for achieving improvements in glycemic control and to use subgroup analyses to elucidate unique characteristics about SMAs that may have been correlated with clinical success. This study may provide valuable information for other facilities considering SMAs.

Overview

The Jesse Brown VAMC (JBVAMC) and the ABJ CBOC implemented a T2DM-focused SMA in 2011. The ABJ CBOC multidisciplinary SMA team consisted of a medical administration service clerk, a registered dietician, a certified DM educator, a registered nurse, a nurse practitioner (NP), and a clinical pharmacy specialist (CPS). This team collaborated to deliver high-quality care to patients with poorly controlled T2DM to improve their glycemic control as well as clinical knowledge of their disease. A private conference room at the ABJ CBOC served as the location for the SMAs. This room was divided into 2 adjacent areas: One area with tables was organized in a semicircle to promote group discussion as well as minimize isolated conversations; the other area had computer terminals to facilitate individualized medication management. Other equipment included a scale for obtaining patient weights and various audio-visual devices.

Related: Efficacy of Patient Aligned Care Team Pharmacist Services in Reaching Diabetes and Hyperlipidemia Treatment Goals

The ABJ CBOC offered monthly SMAs. The team made several attempts to maximize SMA show rates, as previous studies indicated that low SMA show rates were a barrier to success.^3,4,7-9 One review reported no-show rates as high as 70% in certain group visit models.⁴ About 2 weeks prior to a session, prospective SMA patients received automated and customized preappointment letters. Automated and customized phone call reminders were made to prospective SMA patients a few days before each session. As many as 18 patients participated in a single ABJ SMA.

The ABJ SMAs lasted from 60 to 90 minutes, depending on the level of patient participation and the size of the group. The first half of the SMA was dedicated to a group discussion, which involved the SMA team, the patient, and the patient’s family (if desired). The topic of conversation was typically guided by patient curiosity and knowledge deficits in a spontaneous and free-flowing manner; for this reason, these sessions were considered to be open.

The team also engaged in more structured focused sessions, which limited the spontaneous flow of conservation and narrowed the scope to provide targeted education about various aspects of T2DM care. During focused sessions, services such as dental, optometry, podiatry, MOVE! (a VA self-management weight reduction program), and nutrition also participated. Focused sessions addressed topics such as hypoglycemia management, eating around the holidays, sick-day management of T2DM, grocery shopping, exercise, oral health, eye care, and foot care. The specialty services were encouraged to be creative and interactive during the SMA. Many of these services used supportive literature, demonstrations, diagrams, and props to enrich the educational experience. Group discussion typically lasted 30 to 40 minutes; after which patients met individually with either a CPS or NP for medication management.

Medication management focused on optimizing T2DM therapy (both oral and injectable) to improve glycemic control. Interventions outside of T2DM therapy (eg, cholesterol, hypertension, and other risk reduction modalities) were not made, due to time constraints. Once a patient demonstrated improved working knowledge of T2DM and a clinically significant reduction in their glycosylated hemoglobin A_1c (A_1c) they were discharged from SMAs at the discretion of the SMA team. There was no set minimum or maximum duration for the SMAs.

Methods

This study was a retrospective chart review conducted at the JBVAMC and was approved by the institutional review board and the research and development committee. Patient confidentiality was maintained by identifying patients by means other than name or unique identifiers. Protected health information was accessible only by the aforementioned investigators. There was no direct patient contact during this study.

Patient lists were generated from the computerized patient record system (CPRS). Patients were tracked up to 6 months after SMA discharge or until the last SMA in which they participated. The control group was matched according to location, age, glycemic control, and time. The control group never attended an ABJ SMA but may have received regular care through their primary care provider, CPS, or endocrinologist. Prospective control group patients were randomized and reviewed sequentially to obtain the matched cohort.

The study took place at ABJ, an outpatient clinic serving veterans in northwest Indiana and surrounding areas. Inclusion criteria for the SMA group were patients with T2DM, aged ≥ 45 years, with an A_1c ≥ 8.5% seen at ABJ for T2DM from May 1, 2011, to June 30, 2013. The control group included patients with T2DM, aged ≥ 45 years, with an A_1c > 9% who never attended SMAs but may have received regular care at ABJ during the study period. The SMA group’s inclusion criteria threshold for A_1c was lower in order to maximize sample size. The control group’s inclusion criteria threshold for A_1c was higher due to use of a default reminder report called “A_1c > 9%” to generate patient lists. Patients were excluded from the study if they did not meet inclusion criteria.

Baseline datum was the most recent parameter available in CPRS prior to enrollment. The endpoint datum was the parameter nearest the time of SMA discharge or the first available parameter within 6 months from the date of discharge. In the control group, the baseline datum was the initial parameter during the study period and the endpoint datum was the closest measure to 4 months after baseline. Four months was chosen to allow for at least 1 A_1c measurement during the study period. In addition, it was estimated (prior to collecting any data) that 4 months was the average time a patient participated in SMAs. Serial A_1c measurements were defined as values obtained at SMA discharge and 3- and 6-months postdischarge. These parameters were used to evaluate the sustainability of improvements in glycemic control. All values falling outside of these defined parameters were excluded.

Related: Experiences of Veterans With Diabetes From Shared Medical Appointments

The data analysis compared A_1c change from baseline to endpoint for the SMA and control groups. Data collection included baseline characteristics, SMA show rate, number of SMA patients seen by a CPS or NP, number and type of SMA interventions made by a CPS or NP, and the number and type of non-SMA interventions made during the study period. Intervention types were medications: added, discontinued, or titrated; and other, defined as referrals made to specialty services (such as dental, optometry, and podiatry).

Secondary endpoints included the number of SMAs and glycemic improvement, SMA format style (open- vs focused session) and glycemic improvement, SMA provider (CPS vs NP) and glycemic improvement, the change in A_1c stratified by baseline A_1c (A_1c ≥ 10% vs < 10%), the change in actual body weight (ABW) and body mass index (BMI), and maintenance of A_1c (3- and 6-months postdischarge).

The primary endpoint was evaluated using a 2-sample Student t test. Secondary endpoints were evaluated using the independent t test. Statistical significance was defined as P < .05.

Results

A total of 129 unique patients were scheduled for SMAs, 62 of which met inclusion criteria and were included in the SMA group. During enrollment, 67 patients were excluded: 55 never participated in SMAs, 6 had baseline A_1c values < 8.5%, 4 had insufficient data, and 2 were aged < 45 years. A total of 29 SMAs were conducted during the study period, and patients attended an average of 3.15 ± 2.14 (SD) SMAs. The average attendance at each SMA was 7.1 ± 2.62 (SD) patients. For the control group, 754 unique patients were identified and randomized. A total of 90 charts were sequentially reviewed in order to obtain the 62 patients for the control group.

Baseline characteristics were balanced between groups. However, there were more women in the SMA group vs the control group (Table 1). Within the control group, there were a total of 107 appointments that addressed T2DM, which averaged 1.72 ± 1.51 (SD) appointments per patient. The total number of interventions made in the SMA group was 192: 64.6% (124) by a CPS and 35.4% (68) by a NP. For the CPS, the most frequent intervention was medication titration (69.5%), followed by other (23.5%), medication addition (4%), and medication discontinuation (3%). Of note, 53.2% (33) of the SMA patients were seen an average of 1.2 times by non-SMA providers. The SMA patients had a total of 45 non-SMA interventions (0.73 per patient) during the study period.

For the primary endpoint, the SMA group had a 1.48% ± 0.02 (SD) reduction in A_1c compared with a 0.6% ± 0.02 (SD) decrease in the control group (P = .01). When evaluating mean changes in A_1c by the number of SMAs attended, it was noted that participation in ≥ 6 SMAs led to the greatest reduction in A_1c of 2.08%. In the SMA group, it was noted that patients with higher A_1c values at baseline demonstrated greater improvements in glycemic control compared with patients with lower baseline A_1c values. The mean change in A_1c, stratified by baseline A_1c, was -2.26% for those with baseline A_1c values ≥ 10% and -0.87% for those with baseline A_1c values < 10%.

In evaluating the format style, open- vs focused-session, it was observed that participation in focused sessions led to greater improvements in glycemic control. Furthermore, when stratified by provider, greater improvements in glycemic control were demonstrated when medication management was completed by a CPS vs a NP (Table 2). The average number of interventions per SMA patient was 3.1 ± 2.22 (SD). For the control group, the total number of interventions made was 86, with an average of 1.37 ± 1.51 (SD) per patient. The overall show rate was 49% ± 16 (SD), 52% ± 16 (SD) for open visits, and 46% ± 15 (SD) for focused visits. The mean change in ABW and BMI from baseline to endpoint was no different between the SMA and control groups (Table 3). The SMA group participants demonstrated a decrease in A_1c at 3 months postdischarge, and a moderate increase in A_1c was noted at 6 months postdischarge.

Discussion

Shared medical appointments provide an effective alternative to standards of care in order to obtain improvements in glycemic control. Consistent with previous studies, this study reported significant improvements in glycemic control in the SMA group vs the control group. This study also elucidated unique characteristics about SMAs that may have been correlated with clinical success.

Although the greatest improvements in glycemic control were noted for those who participated in ≥ 6 SMAs, it was observed that participation in only 1 SMA also led to improvements. For a site with limited staff and a high volume of patients waiting to participate in SMAs, it may be mutually beneficial to offer only 1 SMA per patient. In addition, patients with ≥ 10% A_1c at baseline demonstrated greater improvements in glycemic control compared to those with < 10% A_1c at baseline. The reasons the higher baseline A_1c subgroup responded to interventions more robustly are unclear and likely multifactorial. Nonetheless, factors such as psychosocial influences (eg, peer pressure to get healthy) may have increased motivation to prevent complications and improved medication adherence in the setting of closer follow-up. Additionally, hyperresponsiveness to drug therapy may have played a role. Regardless, for new SMA programs interested in making an immediate impact, it may be advantageous to initially select patients with very poorly controlled DM.

A unique aspect of the ABJ SMA was the variety of focused sessions offered. Previous studies did not demonstrate such a variety of focused sessions, nor did they evaluate the impact of a focused visit on the patient’s T2DM control. Participation in focused ABJ SMA sessions may have led to improved T2DM control, which may be attributed to the value patients assigned to specialty care and an increased motivation to get healthy.

Related: SGLT2 Inhibitors for Type 2 Diabetes Mellitus Treatment

Another factor that may have led to improved T2DM control was CPS involvement with medication management. The presence of a NP was highly valued, both from a group discussion and medication management standpoint; still, it is a good idea to involve a CPS who has a strong command of DM pharmacotherapy. One shortcoming of this SMA program was the inability for patients to maintain glycemic improvements 6 months after discharge. This pitfall was likely the result of suboptimal coordination of care after SMA discharge and may be avoided by asking the medical administration service clerk to promptly schedule discharged SMA patients for a general medicine clinic T2DM follow-up.The SMA patients had more T2DM interventions within the same time frame compared with the control patients. Although not causative, the increased number of interventions in addition to the bolstered support of the SMA may have correlated with glycemic improvements.

An important finding of this study was the SMA show rate and how it compared with attendance rates found in other group models. The favorable ABJ SMA show rate could have been due to the rigorous attention paid to reminder letters and phone calls. The literature has not established a standard approach to increasing SMA show rates; however, the current data suggest that increased reminders may have increased attendance.

Limitations

This study had several limitations. The external validity was weakened by the modest sample size and the homogenous baseline characteristics of those enrolled. Another limitation was inconsistent documentation of laboratory parameters. The inability to obtain A_1c values exactly at enrollment and discharge could have potentially skewed the results. In addition, incomplete documentation of interventions for dual-care patients (ie, those who obtained care outside of the VA) was an unavoidable challenge. Last, this study did not perform an assessment of SMA patient satisfaction, cost-benefit, or safety.

Conclusion

The ABJ SMA was an effective addition to standards of care in order to achieve improvements in glycemic control in a veteran population with poorly controlled T2DM. Furthermore,the data suggest that a successful program should be multidisciplinary, select poorly controlled patients, offer focused sessions, have a CPS participate in medication management, and encourage patients to complete ≥ 6 sessions. Future studies should be conducted to include more diverse patients to see whether the efficacy of this SMA format is maintained.

A safety analysis should also be conducted to ensure that the SMA format is not only effective, but also a safe means to manage medical conditions. In addition, the scope of the ABJ SMAs should be expanded to allow for evaluation of other diseases. An evaluation of patient satisfaction and cost-benefit could provide additional support for the implementation of SMAs, as improvements in quality of life and cost savings are endpoints to be desired.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

In 2012, 9.3% of the U.S. population had diabetes mellitus (DM).¹ According to the American Diabetes Association, in 2012, the total cost of diagnosed DM in the U.S. was $245 billion.² Diabetes mellitus is a leading cause of blindness, end-stage renal disease, and amputation in the U.S.³ Up to 80% of patients with DM will develop or die of macrovascular disease, such as heart attack or stroke.³

Diabetes mellitus is a chronic disease of epidemic proportion with management complexity that threatens to overwhelm providers in the acute care and primary care settings. Limited specialist availability and increased wait times continue to afflict the VA health care system, prompting efforts to increase health care provider (HCP) access and improve clinic efficiency.⁴ One of the methods proposed to increase HCP access and maximize clinic efficiency is the shared medical appointment (SMA).^5,6

The SMA was designed to improve access and quality of care through enhanced education and support. With the number of people living with chronic diseases on the rise, the current patient-provider model is unrealistic in today’s health care environment. Shared medical appointments offer a unique format for providing evidence-based chronic disease management in which patients and a multidisciplinary team of providers collaborate toward education, discussion, and medication management in a supportive environment.⁷ Research has suggested that SMAs are a successful way to manage type 2 DM (T2DM).^8,9 However, there is uncertainty regarding the optimal model design. The goals of this study were to evaluate whether the diabetes SMA at the Adam Benjamin, Jr. (ABJ) community-based outpatient clinic (CBOC) was an effective practice model for achieving improvements in glycemic control and to use subgroup analyses to elucidate unique characteristics about SMAs that may have been correlated with clinical success. This study may provide valuable information for other facilities considering SMAs.

Overview

The Jesse Brown VAMC (JBVAMC) and the ABJ CBOC implemented a T2DM-focused SMA in 2011. The ABJ CBOC multidisciplinary SMA team consisted of a medical administration service clerk, a registered dietician, a certified DM educator, a registered nurse, a nurse practitioner (NP), and a clinical pharmacy specialist (CPS). This team collaborated to deliver high-quality care to patients with poorly controlled T2DM to improve their glycemic control as well as clinical knowledge of their disease. A private conference room at the ABJ CBOC served as the location for the SMAs. This room was divided into 2 adjacent areas: One area with tables was organized in a semicircle to promote group discussion as well as minimize isolated conversations; the other area had computer terminals to facilitate individualized medication management. Other equipment included a scale for obtaining patient weights and various audio-visual devices.

Related: Efficacy of Patient Aligned Care Team Pharmacist Services in Reaching Diabetes and Hyperlipidemia Treatment Goals

The ABJ CBOC offered monthly SMAs. The team made several attempts to maximize SMA show rates, as previous studies indicated that low SMA show rates were a barrier to success.^3,4,7-9 One review reported no-show rates as high as 70% in certain group visit models.⁴ About 2 weeks prior to a session, prospective SMA patients received automated and customized preappointment letters. Automated and customized phone call reminders were made to prospective SMA patients a few days before each session. As many as 18 patients participated in a single ABJ SMA.

The ABJ SMAs lasted from 60 to 90 minutes, depending on the level of patient participation and the size of the group. The first half of the SMA was dedicated to a group discussion, which involved the SMA team, the patient, and the patient’s family (if desired). The topic of conversation was typically guided by patient curiosity and knowledge deficits in a spontaneous and free-flowing manner; for this reason, these sessions were considered to be open.

The team also engaged in more structured focused sessions, which limited the spontaneous flow of conservation and narrowed the scope to provide targeted education about various aspects of T2DM care. During focused sessions, services such as dental, optometry, podiatry, MOVE! (a VA self-management weight reduction program), and nutrition also participated. Focused sessions addressed topics such as hypoglycemia management, eating around the holidays, sick-day management of T2DM, grocery shopping, exercise, oral health, eye care, and foot care. The specialty services were encouraged to be creative and interactive during the SMA. Many of these services used supportive literature, demonstrations, diagrams, and props to enrich the educational experience. Group discussion typically lasted 30 to 40 minutes; after which patients met individually with either a CPS or NP for medication management.

Medication management focused on optimizing T2DM therapy (both oral and injectable) to improve glycemic control. Interventions outside of T2DM therapy (eg, cholesterol, hypertension, and other risk reduction modalities) were not made, due to time constraints. Once a patient demonstrated improved working knowledge of T2DM and a clinically significant reduction in their glycosylated hemoglobin A_1c (A_1c) they were discharged from SMAs at the discretion of the SMA team. There was no set minimum or maximum duration for the SMAs.

Methods

This study was a retrospective chart review conducted at the JBVAMC and was approved by the institutional review board and the research and development committee. Patient confidentiality was maintained by identifying patients by means other than name or unique identifiers. Protected health information was accessible only by the aforementioned investigators. There was no direct patient contact during this study.

Patient lists were generated from the computerized patient record system (CPRS). Patients were tracked up to 6 months after SMA discharge or until the last SMA in which they participated. The control group was matched according to location, age, glycemic control, and time. The control group never attended an ABJ SMA but may have received regular care through their primary care provider, CPS, or endocrinologist. Prospective control group patients were randomized and reviewed sequentially to obtain the matched cohort.

The study took place at ABJ, an outpatient clinic serving veterans in northwest Indiana and surrounding areas. Inclusion criteria for the SMA group were patients with T2DM, aged ≥ 45 years, with an A_1c ≥ 8.5% seen at ABJ for T2DM from May 1, 2011, to June 30, 2013. The control group included patients with T2DM, aged ≥ 45 years, with an A_1c > 9% who never attended SMAs but may have received regular care at ABJ during the study period. The SMA group’s inclusion criteria threshold for A_1c was lower in order to maximize sample size. The control group’s inclusion criteria threshold for A_1c was higher due to use of a default reminder report called “A_1c > 9%” to generate patient lists. Patients were excluded from the study if they did not meet inclusion criteria.

Baseline datum was the most recent parameter available in CPRS prior to enrollment. The endpoint datum was the parameter nearest the time of SMA discharge or the first available parameter within 6 months from the date of discharge. In the control group, the baseline datum was the initial parameter during the study period and the endpoint datum was the closest measure to 4 months after baseline. Four months was chosen to allow for at least 1 A_1c measurement during the study period. In addition, it was estimated (prior to collecting any data) that 4 months was the average time a patient participated in SMAs. Serial A_1c measurements were defined as values obtained at SMA discharge and 3- and 6-months postdischarge. These parameters were used to evaluate the sustainability of improvements in glycemic control. All values falling outside of these defined parameters were excluded.

Related: Experiences of Veterans With Diabetes From Shared Medical Appointments

The data analysis compared A_1c change from baseline to endpoint for the SMA and control groups. Data collection included baseline characteristics, SMA show rate, number of SMA patients seen by a CPS or NP, number and type of SMA interventions made by a CPS or NP, and the number and type of non-SMA interventions made during the study period. Intervention types were medications: added, discontinued, or titrated; and other, defined as referrals made to specialty services (such as dental, optometry, and podiatry).

Secondary endpoints included the number of SMAs and glycemic improvement, SMA format style (open- vs focused session) and glycemic improvement, SMA provider (CPS vs NP) and glycemic improvement, the change in A_1c stratified by baseline A_1c (A_1c ≥ 10% vs < 10%), the change in actual body weight (ABW) and body mass index (BMI), and maintenance of A_1c (3- and 6-months postdischarge).

The primary endpoint was evaluated using a 2-sample Student t test. Secondary endpoints were evaluated using the independent t test. Statistical significance was defined as P < .05.

Results

A total of 129 unique patients were scheduled for SMAs, 62 of which met inclusion criteria and were included in the SMA group. During enrollment, 67 patients were excluded: 55 never participated in SMAs, 6 had baseline A_1c values < 8.5%, 4 had insufficient data, and 2 were aged < 45 years. A total of 29 SMAs were conducted during the study period, and patients attended an average of 3.15 ± 2.14 (SD) SMAs. The average attendance at each SMA was 7.1 ± 2.62 (SD) patients. For the control group, 754 unique patients were identified and randomized. A total of 90 charts were sequentially reviewed in order to obtain the 62 patients for the control group.

Baseline characteristics were balanced between groups. However, there were more women in the SMA group vs the control group (Table 1). Within the control group, there were a total of 107 appointments that addressed T2DM, which averaged 1.72 ± 1.51 (SD) appointments per patient. The total number of interventions made in the SMA group was 192: 64.6% (124) by a CPS and 35.4% (68) by a NP. For the CPS, the most frequent intervention was medication titration (69.5%), followed by other (23.5%), medication addition (4%), and medication discontinuation (3%). Of note, 53.2% (33) of the SMA patients were seen an average of 1.2 times by non-SMA providers. The SMA patients had a total of 45 non-SMA interventions (0.73 per patient) during the study period.

For the primary endpoint, the SMA group had a 1.48% ± 0.02 (SD) reduction in A_1c compared with a 0.6% ± 0.02 (SD) decrease in the control group (P = .01). When evaluating mean changes in A_1c by the number of SMAs attended, it was noted that participation in ≥ 6 SMAs led to the greatest reduction in A_1c of 2.08%. In the SMA group, it was noted that patients with higher A_1c values at baseline demonstrated greater improvements in glycemic control compared with patients with lower baseline A_1c values. The mean change in A_1c, stratified by baseline A_1c, was -2.26% for those with baseline A_1c values ≥ 10% and -0.87% for those with baseline A_1c values < 10%.

In evaluating the format style, open- vs focused-session, it was observed that participation in focused sessions led to greater improvements in glycemic control. Furthermore, when stratified by provider, greater improvements in glycemic control were demonstrated when medication management was completed by a CPS vs a NP (Table 2). The average number of interventions per SMA patient was 3.1 ± 2.22 (SD). For the control group, the total number of interventions made was 86, with an average of 1.37 ± 1.51 (SD) per patient. The overall show rate was 49% ± 16 (SD), 52% ± 16 (SD) for open visits, and 46% ± 15 (SD) for focused visits. The mean change in ABW and BMI from baseline to endpoint was no different between the SMA and control groups (Table 3). The SMA group participants demonstrated a decrease in A_1c at 3 months postdischarge, and a moderate increase in A_1c was noted at 6 months postdischarge.

Discussion

Shared medical appointments provide an effective alternative to standards of care in order to obtain improvements in glycemic control. Consistent with previous studies, this study reported significant improvements in glycemic control in the SMA group vs the control group. This study also elucidated unique characteristics about SMAs that may have been correlated with clinical success.

Although the greatest improvements in glycemic control were noted for those who participated in ≥ 6 SMAs, it was observed that participation in only 1 SMA also led to improvements. For a site with limited staff and a high volume of patients waiting to participate in SMAs, it may be mutually beneficial to offer only 1 SMA per patient. In addition, patients with ≥ 10% A_1c at baseline demonstrated greater improvements in glycemic control compared to those with < 10% A_1c at baseline. The reasons the higher baseline A_1c subgroup responded to interventions more robustly are unclear and likely multifactorial. Nonetheless, factors such as psychosocial influences (eg, peer pressure to get healthy) may have increased motivation to prevent complications and improved medication adherence in the setting of closer follow-up. Additionally, hyperresponsiveness to drug therapy may have played a role. Regardless, for new SMA programs interested in making an immediate impact, it may be advantageous to initially select patients with very poorly controlled DM.

A unique aspect of the ABJ SMA was the variety of focused sessions offered. Previous studies did not demonstrate such a variety of focused sessions, nor did they evaluate the impact of a focused visit on the patient’s T2DM control. Participation in focused ABJ SMA sessions may have led to improved T2DM control, which may be attributed to the value patients assigned to specialty care and an increased motivation to get healthy.

Related: SGLT2 Inhibitors for Type 2 Diabetes Mellitus Treatment

Another factor that may have led to improved T2DM control was CPS involvement with medication management. The presence of a NP was highly valued, both from a group discussion and medication management standpoint; still, it is a good idea to involve a CPS who has a strong command of DM pharmacotherapy. One shortcoming of this SMA program was the inability for patients to maintain glycemic improvements 6 months after discharge. This pitfall was likely the result of suboptimal coordination of care after SMA discharge and may be avoided by asking the medical administration service clerk to promptly schedule discharged SMA patients for a general medicine clinic T2DM follow-up.The SMA patients had more T2DM interventions within the same time frame compared with the control patients. Although not causative, the increased number of interventions in addition to the bolstered support of the SMA may have correlated with glycemic improvements.

An important finding of this study was the SMA show rate and how it compared with attendance rates found in other group models. The favorable ABJ SMA show rate could have been due to the rigorous attention paid to reminder letters and phone calls. The literature has not established a standard approach to increasing SMA show rates; however, the current data suggest that increased reminders may have increased attendance.

Limitations

This study had several limitations. The external validity was weakened by the modest sample size and the homogenous baseline characteristics of those enrolled. Another limitation was inconsistent documentation of laboratory parameters. The inability to obtain A_1c values exactly at enrollment and discharge could have potentially skewed the results. In addition, incomplete documentation of interventions for dual-care patients (ie, those who obtained care outside of the VA) was an unavoidable challenge. Last, this study did not perform an assessment of SMA patient satisfaction, cost-benefit, or safety.

Conclusion

The ABJ SMA was an effective addition to standards of care in order to achieve improvements in glycemic control in a veteran population with poorly controlled T2DM. Furthermore,the data suggest that a successful program should be multidisciplinary, select poorly controlled patients, offer focused sessions, have a CPS participate in medication management, and encourage patients to complete ≥ 6 sessions. Future studies should be conducted to include more diverse patients to see whether the efficacy of this SMA format is maintained.

A safety analysis should also be conducted to ensure that the SMA format is not only effective, but also a safe means to manage medical conditions. In addition, the scope of the ABJ SMAs should be expanded to allow for evaluation of other diseases. An evaluation of patient satisfaction and cost-benefit could provide additional support for the implementation of SMAs, as improvements in quality of life and cost savings are endpoints to be desired.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Atopic dermatitis (AD) is a chronic inflammatory skin disease that typically begins in early childhood (Figure). It is one of the most commonly diagnosed dermatologic conditions, affecting up to 25% of children and 2% to 3% of adults in the United States.^1,2 The mainstays of treatment for AD are topical emollients and topical medications, of which corticosteroids are most commonly prescribed.³ Although treatments for AD generally are straightforward and efficacious when used correctly, poor adherence to treatment often prevents patients from achieving disease control.⁴ Patient adherence to therapy is a familiar challenge in dermatology, especially for diseases like AD that require long-term treatment with topical medications.^4,5 In some instances, poor adherence may be misconstrued as poor response to treatment, which may lead to escalation to more powerful and potentially dangerous systemic medications.⁶ Ensuring good adherence to treatment leads to better outcomes and disease control, averts unnecessary treatment, prevents disease complications, improves quality of life, and decreases treatment cost.^4,5 This article provides a review of the literature on patient adherence to topical therapies for AD as well as a discussion of methods to improve patient adherence to treatment in the clinical setting.

Methods

A PubMed search of articles indexed for MEDLINE from January 2005 to May 2015 was conducted to identify studies that focused on treatment adherence in AD using the search terms atopic dermatitis and medication adherence and atopic dermatitis and patient compliance After excluding duplicate results and those that were not in the English language, a final list of clinical trials that investigated patient adherence/compliance to topical medications for the treatment of AD was extracted for evaluation.

Results

Our review of the literature yielded 7 quantitative studies that evaluated adherence to topical medications in AD using electronic monitoring and/or self-reporting (Table).^7-13 Participant demographics, disease severity, drug and vehicle used, duration of treatment, and number of follow-up visits varied. All studies used medication event monitoring system caps on medication jars to objectively track patient adherence by recording the date and time when the cap was removed. To assess disease response, the studies used such measures as the Investigator Global Assessment scale, Eczema Area and Severity Index score, or other visual analog scales.

In all of the studies, treatment proved effective and disease severity declined from baseline regardless of the rate of adherence, with benefit continuing after treatment had ended.^7-13 Some results suggested that better adherence increased treatment efficacy and reduced disease severity.^8,9 However, one 10-day trial found no difference in severity and efficacy among participants who applied the medication at least once daily, missed applications some days, or applied the medication more than twice daily.¹³

Study participants typically overestimated their adherence to treatment compared to actual adherence rates, with most reporting near 100% adherence.^7-9,11,12 Average measured adherence rates ranged from 32% to 93% (Table). Adherence rates typically were highest at the beginning of the study and decreased as the study continued.^7-13 The study with the best average adherence rate of 93% had the shortest treatment period of 3 days,¹¹ and the study with the lowest average adherence rate of 32% had the longest treatment period of 8 weeks.⁷ The study with the lowest adherence rate was the only study wherein participants were blinded to their enrollment in the study, which would most closely mimic adherence rates in clinical practice.⁷ The participants in the other studies were not aware that their adherence was being monitored, but their behavior may have been influenced since they were aware of their enrollment in the study.

Many variables affect treatment adherence in patients with AD. Average adherence rates were significantly higher (P=.03) in participants with greater disease severity.⁷ There is conflicting evidence regarding the role of medication vehicle in treatment adherence. While Wilson et al⁹ did not find any difference in adherence based on medication vehicle, Yentzer et al¹² found vehicle characteristics and medication side effects were among patients’ top-ranked concerns about using topical medications. Sagransky et al¹⁰ compared treatment adherence between 2 groups of AD patients: one control group received a standard-of-care 4-week follow-up, and an active group received an additional 1-week follow-up. The mean adherence rate of the treatment group was 69% compared with 54% in the control group.¹⁰

Comment

Poor adherence to treatment is a pervasive problem in patients with AD. Our review of the literature confirmed that patients generally are not accurate historians of their medication usage, often reporting near-perfect treatment adherence even when actual adherence is poor. Rates of adherence from clinical trials are likely higher than those seen in clinical practice due in part to study incentives and differences between how patients in a study are treated compared to those in a physician’s clinic; for example, research study participants often have additional follow-up visits compared to those being treated in the clinical population and by virtue of being enrolled in a study are aware that their behavior is being monitored, which can increase treatment adherence.⁷

The dogma suggesting that tachyphylaxis can occur with long-term use of topical corticosteroids is not supported by clinical trials.¹⁴ Furthermore, in our review of the literature patient adherence was highest in the shortest study¹¹ and lowest in the longest study.⁷ Given that AD patients cannot benefit from a treatment if they do not use it, the supposed decrease in efficacy of topical corticosteroids over time may be because patients fail to use them consistently.

Our review of the literature was limited by the small body of research that exists on treatment adherence in AD patients, especially relating to topical medications, and did not reveal any studies evaluating systemic medications in AD. Of the studies we examined, sample sizes were small and treatment and follow-up periods were short. Our review only covered adherence to prescribed topical medications in AD, chiefly corticosteroids; thus, we did not evaluate adherence to other therapies (eg, emollients) in this patient population.

The existing research also is limited by the relative paucity of data showing a correlation between improved adherence to topical treatment and improved disease outcomes, which may be due to the methodological limitations of the study designs that have been used; for instance, studies may use objective monitors to describe daily adherence to treatment, but disease severity typically is measured over longer periods of time, usually every few weeks or months. Short-term data may not be an accurate demonstration of how participants’ actual treatment adherence impacts disease outcome, as the data does not account for more complex adherence factors; for example, participants who achieve good disease control using topical corticosteroids for an 8-week study period may actually demonstrate poor treatment adherence overall, as topical corticosteroids have good short-term efficacy and the patient may have stopped using the product after the first few weeks of the treatment period. In contrast, poorly adherent patients may never use the medication well enough to achieve improvement and may continue low-level use throughout the study period. Therefore, studies that measure disease severity at more regular intervals are required to show the true effect of treatment adherence on disease outcomes.

Since AD mainly affects children, family issues can pose special challenges to attaining good treatment adherence.^15,16 The physician–patient (or parent) relationship and the family’s perception of the patient’s disease severity are strong predictors of adherence to topical treatment.¹⁶ Potential barriers to adherence in the pediatric population are caregivers with negative beliefs about treatment, the time-consuming nature of applying topical therapies, or a child who is uncooperative.^15,17 In the treatment of infants, critical factors are caregiver availability and beliefs and fears about medications and their side effects, while in the teenage population, the desire to “fit in” and oppositional behavior can lead to poor adherence to treatment.¹⁷ Regardless of age, other barriers to treatment adherence are forgetfulness, belief that the drug is not working, and the messiness of treatment.¹⁷

Educational tools (eg, action plans, instructions about how to apply topical medications correctly) may be underutilized in patients with AD. If consistently implemented, these tools could have a positive impact on adherence to medication in patients with AD. For example, written action plans pioneered in the asthma community have shown to improve quality of life and reduce disease severity and may offer the same benefits for AD patients due to the similarities of the diseases.¹⁸ Since AD patients and their caregivers often are not well versed in how to apply topical medications correctly, efforts to educate patients could potentially increase adherence to treatment. In one study, AD patients began to use medications more effectively after applying a fluorescent cream to reveal affected areas they had missed, and clinicians were able to provide additional instruction based on the findings.¹⁹

Adherence to topical treatments among AD patients is a multifactorial issue. Regimens often are complex and inconvenient due to the need for multiple medications, the topical nature of the products, and the need for frequent application. To optimize prescription treatments, patients also must be diligent with preventive measures such as application of topical emollients and use of bathing techniques (eg, bleach baths). A way to overcome treatment complexity and increase adherence may be to provide a written action plan and involve the patient and caregiver in the plan’s development. If a drug formulation is not aesthetically acceptable to the patient (eg, the greasiness of an ointment), allowing the patient to choose the medication vehicle may increase satisfaction and use.¹² Fear of steroid side effects also is common among patients and caregivers and could be overcome with education about the product.²⁰

Conclusion

Treatment adherence can have a dramatic effect on diseases outcomes and can be particularly challenging in AD due to the use of topical medications with complex treatment regimens. Additionally, a large majority of patients with AD are children, from infants to teenagers, adding another layer of treatment challenges. Further research is needed to more definitively develop effective methods for enhancing treatment adherence in this patient population. Although enormous amounts of money are being spent to develop improved treatments for AD, we may be able to achieve far more benefit at a much lower cost by figuring out how to get patients to adhere to the treatments that are already available.

Atopic dermatitis (AD) is a chronic inflammatory skin disease that typically begins in early childhood (Figure). It is one of the most commonly diagnosed dermatologic conditions, affecting up to 25% of children and 2% to 3% of adults in the United States.^1,2 The mainstays of treatment for AD are topical emollients and topical medications, of which corticosteroids are most commonly prescribed.³ Although treatments for AD generally are straightforward and efficacious when used correctly, poor adherence to treatment often prevents patients from achieving disease control.⁴ Patient adherence to therapy is a familiar challenge in dermatology, especially for diseases like AD that require long-term treatment with topical medications.^4,5 In some instances, poor adherence may be misconstrued as poor response to treatment, which may lead to escalation to more powerful and potentially dangerous systemic medications.⁶ Ensuring good adherence to treatment leads to better outcomes and disease control, averts unnecessary treatment, prevents disease complications, improves quality of life, and decreases treatment cost.^4,5 This article provides a review of the literature on patient adherence to topical therapies for AD as well as a discussion of methods to improve patient adherence to treatment in the clinical setting.

Methods

A PubMed search of articles indexed for MEDLINE from January 2005 to May 2015 was conducted to identify studies that focused on treatment adherence in AD using the search terms atopic dermatitis and medication adherence and atopic dermatitis and patient compliance After excluding duplicate results and those that were not in the English language, a final list of clinical trials that investigated patient adherence/compliance to topical medications for the treatment of AD was extracted for evaluation.

Results

Our review of the literature yielded 7 quantitative studies that evaluated adherence to topical medications in AD using electronic monitoring and/or self-reporting (Table).^7-13 Participant demographics, disease severity, drug and vehicle used, duration of treatment, and number of follow-up visits varied. All studies used medication event monitoring system caps on medication jars to objectively track patient adherence by recording the date and time when the cap was removed. To assess disease response, the studies used such measures as the Investigator Global Assessment scale, Eczema Area and Severity Index score, or other visual analog scales.

In all of the studies, treatment proved effective and disease severity declined from baseline regardless of the rate of adherence, with benefit continuing after treatment had ended.^7-13 Some results suggested that better adherence increased treatment efficacy and reduced disease severity.^8,9 However, one 10-day trial found no difference in severity and efficacy among participants who applied the medication at least once daily, missed applications some days, or applied the medication more than twice daily.¹³

Study participants typically overestimated their adherence to treatment compared to actual adherence rates, with most reporting near 100% adherence.^7-9,11,12 Average measured adherence rates ranged from 32% to 93% (Table). Adherence rates typically were highest at the beginning of the study and decreased as the study continued.^7-13 The study with the best average adherence rate of 93% had the shortest treatment period of 3 days,¹¹ and the study with the lowest average adherence rate of 32% had the longest treatment period of 8 weeks.⁷ The study with the lowest adherence rate was the only study wherein participants were blinded to their enrollment in the study, which would most closely mimic adherence rates in clinical practice.⁷ The participants in the other studies were not aware that their adherence was being monitored, but their behavior may have been influenced since they were aware of their enrollment in the study.

Many variables affect treatment adherence in patients with AD. Average adherence rates were significantly higher (P=.03) in participants with greater disease severity.⁷ There is conflicting evidence regarding the role of medication vehicle in treatment adherence. While Wilson et al⁹ did not find any difference in adherence based on medication vehicle, Yentzer et al¹² found vehicle characteristics and medication side effects were among patients’ top-ranked concerns about using topical medications. Sagransky et al¹⁰ compared treatment adherence between 2 groups of AD patients: one control group received a standard-of-care 4-week follow-up, and an active group received an additional 1-week follow-up. The mean adherence rate of the treatment group was 69% compared with 54% in the control group.¹⁰

Comment

Poor adherence to treatment is a pervasive problem in patients with AD. Our review of the literature confirmed that patients generally are not accurate historians of their medication usage, often reporting near-perfect treatment adherence even when actual adherence is poor. Rates of adherence from clinical trials are likely higher than those seen in clinical practice due in part to study incentives and differences between how patients in a study are treated compared to those in a physician’s clinic; for example, research study participants often have additional follow-up visits compared to those being treated in the clinical population and by virtue of being enrolled in a study are aware that their behavior is being monitored, which can increase treatment adherence.⁷

The dogma suggesting that tachyphylaxis can occur with long-term use of topical corticosteroids is not supported by clinical trials.¹⁴ Furthermore, in our review of the literature patient adherence was highest in the shortest study¹¹ and lowest in the longest study.⁷ Given that AD patients cannot benefit from a treatment if they do not use it, the supposed decrease in efficacy of topical corticosteroids over time may be because patients fail to use them consistently.

Our review of the literature was limited by the small body of research that exists on treatment adherence in AD patients, especially relating to topical medications, and did not reveal any studies evaluating systemic medications in AD. Of the studies we examined, sample sizes were small and treatment and follow-up periods were short. Our review only covered adherence to prescribed topical medications in AD, chiefly corticosteroids; thus, we did not evaluate adherence to other therapies (eg, emollients) in this patient population.

The existing research also is limited by the relative paucity of data showing a correlation between improved adherence to topical treatment and improved disease outcomes, which may be due to the methodological limitations of the study designs that have been used; for instance, studies may use objective monitors to describe daily adherence to treatment, but disease severity typically is measured over longer periods of time, usually every few weeks or months. Short-term data may not be an accurate demonstration of how participants’ actual treatment adherence impacts disease outcome, as the data does not account for more complex adherence factors; for example, participants who achieve good disease control using topical corticosteroids for an 8-week study period may actually demonstrate poor treatment adherence overall, as topical corticosteroids have good short-term efficacy and the patient may have stopped using the product after the first few weeks of the treatment period. In contrast, poorly adherent patients may never use the medication well enough to achieve improvement and may continue low-level use throughout the study period. Therefore, studies that measure disease severity at more regular intervals are required to show the true effect of treatment adherence on disease outcomes.

Since AD mainly affects children, family issues can pose special challenges to attaining good treatment adherence.^15,16 The physician–patient (or parent) relationship and the family’s perception of the patient’s disease severity are strong predictors of adherence to topical treatment.¹⁶ Potential barriers to adherence in the pediatric population are caregivers with negative beliefs about treatment, the time-consuming nature of applying topical therapies, or a child who is uncooperative.^15,17 In the treatment of infants, critical factors are caregiver availability and beliefs and fears about medications and their side effects, while in the teenage population, the desire to “fit in” and oppositional behavior can lead to poor adherence to treatment.¹⁷ Regardless of age, other barriers to treatment adherence are forgetfulness, belief that the drug is not working, and the messiness of treatment.¹⁷

Educational tools (eg, action plans, instructions about how to apply topical medications correctly) may be underutilized in patients with AD. If consistently implemented, these tools could have a positive impact on adherence to medication in patients with AD. For example, written action plans pioneered in the asthma community have shown to improve quality of life and reduce disease severity and may offer the same benefits for AD patients due to the similarities of the diseases.¹⁸ Since AD patients and their caregivers often are not well versed in how to apply topical medications correctly, efforts to educate patients could potentially increase adherence to treatment. In one study, AD patients began to use medications more effectively after applying a fluorescent cream to reveal affected areas they had missed, and clinicians were able to provide additional instruction based on the findings.¹⁹

Adherence to topical treatments among AD patients is a multifactorial issue. Regimens often are complex and inconvenient due to the need for multiple medications, the topical nature of the products, and the need for frequent application. To optimize prescription treatments, patients also must be diligent with preventive measures such as application of topical emollients and use of bathing techniques (eg, bleach baths). A way to overcome treatment complexity and increase adherence may be to provide a written action plan and involve the patient and caregiver in the plan’s development. If a drug formulation is not aesthetically acceptable to the patient (eg, the greasiness of an ointment), allowing the patient to choose the medication vehicle may increase satisfaction and use.¹² Fear of steroid side effects also is common among patients and caregivers and could be overcome with education about the product.²⁰

Conclusion

Treatment adherence can have a dramatic effect on diseases outcomes and can be particularly challenging in AD due to the use of topical medications with complex treatment regimens. Additionally, a large majority of patients with AD are children, from infants to teenagers, adding another layer of treatment challenges. Further research is needed to more definitively develop effective methods for enhancing treatment adherence in this patient population. Although enormous amounts of money are being spent to develop improved treatments for AD, we may be able to achieve far more benefit at a much lower cost by figuring out how to get patients to adhere to the treatments that are already available.

Atopic dermatitis (AD) is a chronic inflammatory skin disease that typically begins in early childhood (Figure). It is one of the most commonly diagnosed dermatologic conditions, affecting up to 25% of children and 2% to 3% of adults in the United States.^1,2 The mainstays of treatment for AD are topical emollients and topical medications, of which corticosteroids are most commonly prescribed.³ Although treatments for AD generally are straightforward and efficacious when used correctly, poor adherence to treatment often prevents patients from achieving disease control.⁴ Patient adherence to therapy is a familiar challenge in dermatology, especially for diseases like AD that require long-term treatment with topical medications.^4,5 In some instances, poor adherence may be misconstrued as poor response to treatment, which may lead to escalation to more powerful and potentially dangerous systemic medications.⁶ Ensuring good adherence to treatment leads to better outcomes and disease control, averts unnecessary treatment, prevents disease complications, improves quality of life, and decreases treatment cost.^4,5 This article provides a review of the literature on patient adherence to topical therapies for AD as well as a discussion of methods to improve patient adherence to treatment in the clinical setting.

Methods

A PubMed search of articles indexed for MEDLINE from January 2005 to May 2015 was conducted to identify studies that focused on treatment adherence in AD using the search terms atopic dermatitis and medication adherence and atopic dermatitis and patient compliance After excluding duplicate results and those that were not in the English language, a final list of clinical trials that investigated patient adherence/compliance to topical medications for the treatment of AD was extracted for evaluation.

Results

Our review of the literature yielded 7 quantitative studies that evaluated adherence to topical medications in AD using electronic monitoring and/or self-reporting (Table).^7-13 Participant demographics, disease severity, drug and vehicle used, duration of treatment, and number of follow-up visits varied. All studies used medication event monitoring system caps on medication jars to objectively track patient adherence by recording the date and time when the cap was removed. To assess disease response, the studies used such measures as the Investigator Global Assessment scale, Eczema Area and Severity Index score, or other visual analog scales.

In all of the studies, treatment proved effective and disease severity declined from baseline regardless of the rate of adherence, with benefit continuing after treatment had ended.^7-13 Some results suggested that better adherence increased treatment efficacy and reduced disease severity.^8,9 However, one 10-day trial found no difference in severity and efficacy among participants who applied the medication at least once daily, missed applications some days, or applied the medication more than twice daily.¹³

Study participants typically overestimated their adherence to treatment compared to actual adherence rates, with most reporting near 100% adherence.^7-9,11,12 Average measured adherence rates ranged from 32% to 93% (Table). Adherence rates typically were highest at the beginning of the study and decreased as the study continued.^7-13 The study with the best average adherence rate of 93% had the shortest treatment period of 3 days,¹¹ and the study with the lowest average adherence rate of 32% had the longest treatment period of 8 weeks.⁷ The study with the lowest adherence rate was the only study wherein participants were blinded to their enrollment in the study, which would most closely mimic adherence rates in clinical practice.⁷ The participants in the other studies were not aware that their adherence was being monitored, but their behavior may have been influenced since they were aware of their enrollment in the study.

Many variables affect treatment adherence in patients with AD. Average adherence rates were significantly higher (P=.03) in participants with greater disease severity.⁷ There is conflicting evidence regarding the role of medication vehicle in treatment adherence. While Wilson et al⁹ did not find any difference in adherence based on medication vehicle, Yentzer et al¹² found vehicle characteristics and medication side effects were among patients’ top-ranked concerns about using topical medications. Sagransky et al¹⁰ compared treatment adherence between 2 groups of AD patients: one control group received a standard-of-care 4-week follow-up, and an active group received an additional 1-week follow-up. The mean adherence rate of the treatment group was 69% compared with 54% in the control group.¹⁰

Comment

Poor adherence to treatment is a pervasive problem in patients with AD. Our review of the literature confirmed that patients generally are not accurate historians of their medication usage, often reporting near-perfect treatment adherence even when actual adherence is poor. Rates of adherence from clinical trials are likely higher than those seen in clinical practice due in part to study incentives and differences between how patients in a study are treated compared to those in a physician’s clinic; for example, research study participants often have additional follow-up visits compared to those being treated in the clinical population and by virtue of being enrolled in a study are aware that their behavior is being monitored, which can increase treatment adherence.⁷

The dogma suggesting that tachyphylaxis can occur with long-term use of topical corticosteroids is not supported by clinical trials.¹⁴ Furthermore, in our review of the literature patient adherence was highest in the shortest study¹¹ and lowest in the longest study.⁷ Given that AD patients cannot benefit from a treatment if they do not use it, the supposed decrease in efficacy of topical corticosteroids over time may be because patients fail to use them consistently.

Our review of the literature was limited by the small body of research that exists on treatment adherence in AD patients, especially relating to topical medications, and did not reveal any studies evaluating systemic medications in AD. Of the studies we examined, sample sizes were small and treatment and follow-up periods were short. Our review only covered adherence to prescribed topical medications in AD, chiefly corticosteroids; thus, we did not evaluate adherence to other therapies (eg, emollients) in this patient population.

The existing research also is limited by the relative paucity of data showing a correlation between improved adherence to topical treatment and improved disease outcomes, which may be due to the methodological limitations of the study designs that have been used; for instance, studies may use objective monitors to describe daily adherence to treatment, but disease severity typically is measured over longer periods of time, usually every few weeks or months. Short-term data may not be an accurate demonstration of how participants’ actual treatment adherence impacts disease outcome, as the data does not account for more complex adherence factors; for example, participants who achieve good disease control using topical corticosteroids for an 8-week study period may actually demonstrate poor treatment adherence overall, as topical corticosteroids have good short-term efficacy and the patient may have stopped using the product after the first few weeks of the treatment period. In contrast, poorly adherent patients may never use the medication well enough to achieve improvement and may continue low-level use throughout the study period. Therefore, studies that measure disease severity at more regular intervals are required to show the true effect of treatment adherence on disease outcomes.

Since AD mainly affects children, family issues can pose special challenges to attaining good treatment adherence.^15,16 The physician–patient (or parent) relationship and the family’s perception of the patient’s disease severity are strong predictors of adherence to topical treatment.¹⁶ Potential barriers to adherence in the pediatric population are caregivers with negative beliefs about treatment, the time-consuming nature of applying topical therapies, or a child who is uncooperative.^15,17 In the treatment of infants, critical factors are caregiver availability and beliefs and fears about medications and their side effects, while in the teenage population, the desire to “fit in” and oppositional behavior can lead to poor adherence to treatment.¹⁷ Regardless of age, other barriers to treatment adherence are forgetfulness, belief that the drug is not working, and the messiness of treatment.¹⁷

Educational tools (eg, action plans, instructions about how to apply topical medications correctly) may be underutilized in patients with AD. If consistently implemented, these tools could have a positive impact on adherence to medication in patients with AD. For example, written action plans pioneered in the asthma community have shown to improve quality of life and reduce disease severity and may offer the same benefits for AD patients due to the similarities of the diseases.¹⁸ Since AD patients and their caregivers often are not well versed in how to apply topical medications correctly, efforts to educate patients could potentially increase adherence to treatment. In one study, AD patients began to use medications more effectively after applying a fluorescent cream to reveal affected areas they had missed, and clinicians were able to provide additional instruction based on the findings.¹⁹

Adherence to topical treatments among AD patients is a multifactorial issue. Regimens often are complex and inconvenient due to the need for multiple medications, the topical nature of the products, and the need for frequent application. To optimize prescription treatments, patients also must be diligent with preventive measures such as application of topical emollients and use of bathing techniques (eg, bleach baths). A way to overcome treatment complexity and increase adherence may be to provide a written action plan and involve the patient and caregiver in the plan’s development. If a drug formulation is not aesthetically acceptable to the patient (eg, the greasiness of an ointment), allowing the patient to choose the medication vehicle may increase satisfaction and use.¹² Fear of steroid side effects also is common among patients and caregivers and could be overcome with education about the product.²⁰

Conclusion

Treatment adherence can have a dramatic effect on diseases outcomes and can be particularly challenging in AD due to the use of topical medications with complex treatment regimens. Additionally, a large majority of patients with AD are children, from infants to teenagers, adding another layer of treatment challenges. Further research is needed to more definitively develop effective methods for enhancing treatment adherence in this patient population. Although enormous amounts of money are being spent to develop improved treatments for AD, we may be able to achieve far more benefit at a much lower cost by figuring out how to get patients to adhere to the treatments that are already available.