An Effective and Efficient Testing Protocol for Diagnosing Iron-deficiency Anemia Preoperatively

We obtained permission from the Institutional Review Board at Northwestern University (Chicago, Illinois) to collect and analyze retrospective data from patients seen in the preoperative clinic. This study (Northwestern University Institutional Review Board identification number STU00207504) was granted a waiver of consent because there was no direct patient contact, and all patient data were deidentified. Patient records were queried from the Northwestern Medicine Enterprise Data Warehouse, a centralized and secure database repository for patient information. All patients (ages 16 to 89 yr of age) with completed appointments at the preoperative clinic between January 1, 2015, and May 15, 2016, and between January 1, 2017, and May 15, 2018, who were scheduled for either surgery or procedures requiring anesthesia and who had a complete blood count or a preprocedure anemia evaluation ordered at the time of the clinic appointment were included in the study. Demographic information, including age, sex, race, body mass index, and American Society of Anesthesiologists (ASA) Physical Status classification, were collected. We also downloaded laboratory data including hemoglobin concentration, hematocrit, serum iron, serum ferritin, total iron-binding capacity, transferrin saturation, reticulocyte count, creatinine concentration, estimated glomerular filtration rate, thyroid-stimulating hormone and vitamin B₁₂ concentrations. Once downloaded from the Enterprise Data Warehouse, all data were deidentified. This is the primary analysis of these data.

Between January 2015 and May 2016, patients (ages 16 to 89 yr) scheduled for surgery with completed appointments at the preoperative clinic were screened for preoperative anemia based on established clinic criteria using the traditional approach of indicated testing with follow-up at the discretion of the ordering provider. A complete blood count was ordered at the discretion of the clinicians or the surgical service. The preoperative clinic has guidelines that recommend complete blood count testing for patients based on the potential for significant blood loss, the planned procedure, a history of anemia, alcohol abuse, bleeding, liver or kidney disease, dyspnea, hematologic disorders, malignancy, inflammatory bowel disease, body mass index less than 16 kg/m², malnutrition, palpitations, or syncope.

The preprocedure anemia evaluation was implemented by the preoperative clinic on January 1, 2017. With the preprocedure anemia evaluation order, the phlebotomist is directed to obtain two vials of blood during the initial preoperative clinic visit. For complete blood count testing, the patient’s blood is placed into a lavender-top blood tube containing EDTA, an anticoagulant used for most hematology tests. For additional testing for serum iron concentration, serum ferritin concentration, total iron-binding capacity, transferrin saturation, creatinine concentration, estimated glomerular filtration rate, thyroid-stimulating hormone concentration, and vitamin B₁₂ concentration, the blood is collected in a green-top blood tube containing heparin. First, the lavender-top blood tube is routed to the hematology testing department, and based on the results of the complete blood count, a second order is automatically initiated to test the second vial of blood, a green-top heparinized blood tube, in the chemistry diagnostic department (fig. 1). The World Health Organization (Geneva, Switzerland) defines anemia as a hemoglobin less than 12 g/dl for women and less than 13 g/dl for men. We set the hemoglobin trigger for secondary testing at 12 g/dl or lower to simplify the process. A reticulocyte count is done using the original lavender-top blood tube that had been used for the initial complete blood count. If the hemoglobin concentration is higher than 12 g/dl, the green-top blood tube is discarded, and the patient is not charged for the secondary set of laboratory tests.

Because the preprocedure anemia evaluation was introduced as an option for evaluating anemia in the preoperative period, it was ordered at the discretion of the provider. The traditional approach of ordering a complete blood count was still available. As such, in addition to patients being tested with the preprocedure anemia evaluation, some patients were screened for preoperative anemia using the traditional approach with complete blood count, with follow-up diagnostic anemia testing at the discretion of the ordering provider. The preoperative clinic guidelines for recommending preoperative testing of hemoglobin concentration remained the same after the implementation of the preprocedure anemia evaluation option. During the period of January 1, 2017, to May 15, 2018, we collected data for all patients who were tested using either the traditional approach (complete blood count with further testing for diagnosing the type of anemia at the discretion of the provider) or the preprocedure anemia evaluation.

We chose patients screened in the period between January 2015 and May 2016 as a control group. Patients seen in the preoperative clinic during this period were compared with patients seen in the preoperative clinic between January 2017 and May 2018. The data were analyzed from January 1, 2015, to May 15, 2016, and January 1, 2017, to May 15, 2018. This allowed for 17 months of data before preprocedure anemia evaluation implementation and 17 months of data after implementation of the preprocedure anemia evaluation.

Our primary outcome was the proportion of patients seen in the preoperative clinic who were diagnosed with iron-deficiency anemia using the preprocedure anemia evaluation (January 2017 to May 2018) compared with those diagnosed with iron-deficiency anemia using traditional complete blood count testing before (January 2015 to May 2016) and after (January 2017 to May 2018) implementation of the preprocedure anemia evaluation. Our secondary outcome was the proportion of patients diagnosed with anemia who were subsequently diagnosed with iron-deficiency anemia using the preprocedure anemia evaluation compared with those diagnosed with iron-deficiency anemia using traditional complete blood count testing both before and after implementation of the preprocedure anemia evaluation. The goal of the secondary outcome was to compare the effectiveness of the preprocedure anemia evaluation with that of the traditional anemia evaluation in identifying iron deficiency once anemia is diagnosed. A diagnosis of iron-deficiency anemia is made when a patient with anemia has a serum ferritin concentration less than 30 ng/ml or transferrin saturation less than 20%.¹³  Once a diagnosis of iron-deficiency anemia is made, follow-up evaluations are done in collaboration with other specialized providers.

No statistical power calculation was conducted before the study. The sample size was based on the available data in the selected time periods. All continuous data (e.g., age, body mass index) were found to be nonnormally distributed by the Shapiro–Wilk W test and so are reported as medians (interquartile range). Ordinal data were reduced to nominal data (e.g., ASA Physical Status) and are reported as number (percentage) of patients, as were the nominal data (e.g., sex). In our preplanned analysis of these data, the hypothesis that the preprocedure anemia evaluation will allow for diagnosis of iron-deficiency anemia in a larger proportion of patients compared with the traditional approach was tested first using a 2 × 3 chi-square test with a two-tailed P < 0.05 as the criterion for rejection of the null hypothesis followed by three 2 × 2 chi-square post hoc tests with a two-tailed P < 0.0167 as the criterion for rejection of the null hypothesis for each test application (StatsDirect version 3.1.22, December 21, 2018, United Kingdom). Secondary hypotheses were similarly tested.

Given the time series nature of the delivery of the preprocedural anemia evaluation, we conducted a post hoc grouped linear regression analysis (StatsDirect) of the proportion of patients diagnosed with iron-deficiency anemia per month by traditional anemia evaluation and by preprocedure anemia evaluation. First, we did an analysis of the proportion of patients diagnosed with iron-deficiency anemia per month by traditional anemia evaluation in 2015 to 2016 and by preprocedure anemia evaluation in 2017 to 2018. Next, we analyzed the proportion of patients diagnosed with iron-deficiency anemia per month by traditional anemia evaluation in 2017 to 2018 and by preprocedure anemia evaluation in 2017 to 2018. Finally, we analyzed the proportion of all patients diagnosed with iron-deficiency anemia per month by traditional anemia evaluation and by preprocedure anemia evaluation. The criterion for rejection of the null hypothesis was a two-tailed P < 0.05.

A total of 38,027 patients scheduled for surgery or procedures requiring anesthesia were seen in the preoperative clinic during both time periods. Patient characteristics and laboratory measurements of the three groups were similar (table 1). The occasionally unavailable patient characteristic (e.g., body mass index) data were assumed to be missing at random. Indeterminate follow-up results were not included in the summarized data. No effort was made to impute missing data.

Before the preprocedure anemia evaluation implementation, from January 1, 2016, to May 15, 2017, 17,902 patients completed appointments, of which 49% (8,816 of 17,902) were screened using the traditional preoperative clinic approach of ordering a complete blood count; 21% (1,826 of 8,816) of those patients were anemic (defined as hemoglobin less than or equal to 12 g/dl). Six percent (107 of 1,826) of those with anemia and 1.2% (107 of 8,816) of the screened patients were diagnosed with iron-deficiency anemia. Of the anemic patients, 1.3% (24 of 1,826) did not have iron-deficiency anemia, and 93% (1,695 of 1,826) did not have follow-up anemia testing.

After the preprocedure anemia evaluation was implemented, 20,125 patients had preoperative clinic appointments during the investigative period of January 1, 2017 to May 15, 2018, of which 23% (4,629 of 20,125) were screened using the traditional complete blood count method and 14% (2,828 of 20,125) were screened using the preprocedure anemia evaluation. Of the 4,629 patients tested using the traditional approach, 17% (781 of 4,629) were anemic. Of the 781 anemic patients, 20% (154 of 781) were diagnosed with iron-deficiency anemia, 7% (59 of 781) did not have iron-deficiency anemia, and 73% (568 of 781) did not have follow-up diagnostic anemia testing. Of the 2,828 patients tested using the preprocedure anemia evaluation, 37% (1,055 of 2,828) were diagnosed with anemia, and 100% had reflex iron studies. Of this cohort of 1,055 anemic patients, 70% (738 of 1,055) were diagnosed with iron-deficiency anemia, 25% (262 of 1,055) did not have iron-deficiency anemia, and 5% (55 of 1,055) had iron studies (ferritin and transferrin saturation values) that were indeterminate (fig. 2).

Using the preprocedure anemia evaluation testing protocol in the preoperative clinic resulted in 95% (1,000 of 1,055) of patients who had a hemoglobin less than or equal to 12 g/dl with complete diagnostic follow-up anemia evaluations. This compares to only 7% (131 of 1,826) of anemic patients who had their anemia diagnosed with traditional complete blood count testing before the preprocedure anemia evaluation implementation having appropriate diagnostic anemia testing. With traditional testing after implementation of the preprocedure anemia evaluation, only 27% (213 of 781) of patients had complete diagnostic anemia testing.

The preprocedure anemia evaluation identified a far larger proportion of patients with iron-deficiency anemia (26.1%, 738 of 2,828) than did the traditional complete blood count before implementation of preprocedure anemia evaluation (107 of 8,816 = 1.2%; relative risk [95% CI], 21.5 [17.6 to 26.2]; P < 0.0001) and after its implementation (154 of 4,629 = 3.3%; relative risk [95% CI], 7.8 [6.6 to 9.3]; P < 0.0001; table 2). The proportion of patients identified with iron-deficiency anemia with traditional complete blood count in 2017 to 2018 was 2.7 times that identified in 2015 to 2016 (95% CI, 2.1 to 3.5; P < 0.0001).

In patients diagnosed with anemia, the preprocedure anemia evaluation identified a larger proportion of patients with iron-deficiency anemia compared with the proportion identified by screening with traditional complete blood count in both time periods (table 3). In addition, the preprocedure anemia evaluation identified a larger proportion of patients with anemia of any type compared with the proportion identified by screening with traditional complete blood count in both time periods (table 4).

The grouped linear regression analysis of the proportion of patients diagnosed with iron deficiency per month by traditional and preprocedure anemia evaluations indicated that the proportion of patients diagnosed with the preprocedure anemia evaluation diverges markedly from the trend projected by the traditional evaluation data (table 5 and fig. 3). Specifically, the vertical separation (i.e., the difference of the Y intercepts of between the proportion of patients diagnosed with iron-deficiency anemia per month by the two methods vs. the months the data were collected) was −0.227 (95% CI, −0.247 to −0.206), consistent with the overall difference in the proportion of patients diagnosed with iron-deficiency anemia by the two methods (table 2).

The preprocedure anemia evaluation identified a larger proportion of patients with iron-deficiency anemia (26.1%) than did screening with traditional complete blood count with follow-up testing, both before implementation of the preprocedure anemia evaluation (1.2%) and after its implementation (3.3%). The grouped linear regression analysis of the proportion of patients diagnosed with iron-deficiency anemia per month by traditional and preprocedure anemia evaluations indicated that the proportion of patients diagnosed with the traditional method in 2017 to 2018 is a natural progression of the trend observed before implementation of the preprocedure anemia evaluation (2015 to 2016), but the proportion diagnosed per month with the preprocedure anemia evaluation diverges markedly from the trend projected by the traditional evaluation data.

Although reflex testing to diagnose anemia has been proposed,^14,15  implementation of such reflex anemia testing has not been described. Typical automated reflex laboratory testing follows a simple algorithm by which the same vial of blood is reanalyzed within the same medical laboratory department for further diagnostics when the initial value is abnormal. The preprocedure anemia evaluation presents a challenge because it requires two different laboratory departments and two different types of blood tubes with different additives designed to stabilize and preserve the specimen for analytical testing. The logistics involved in actualizing the design of the preprocedure anemia evaluation order is outlined under Materials and Methods and illustrated in figure 1.

The process with the preprocedure anemia evaluation differs from the traditional approach of anemia evaluation, which involves multiple patient visits for blood draws and laboratory testing to first determine the presence of anemia and then to diagnose the type of anemia. Preprocedure anemia evaluation reflex testing eliminates the need for return visits for further blood draws and laboratory testing. In addition, it does not require the provider to act on the results of the complete blood count. The preprocedure anemia evaluation order is placed once, and the entire process of reflex testing is automated. The diagnostic tests in the preprocedure anemia evaluation evaluates for both iron-deficiency anemia and vitamin B₁₂-deficiency anemia, the most common, easily treatable types of anemia. It also tests for common medical conditions causing anemia, such as chronic kidney disease and hypothyroidism. If the reflex tests are normal, it is highly likely that anemia of chronic disease (also known as anemia of inflammation) or a congenital hemoglobinopathy are the reasons for low hemoglobin in this preoperative population.

One of the authors (B.J.S.) worked extensively with the information technology department and the clinical laboratory to create the preprocedure anemia evaluation process. The design and implementation involved personnel resources but did not require additional expenditures, such as laboratory equipment. Reproducing this testing is readily feasible but is dependent on functionality of laboratories and information services at other institutions. The preprocedure anemia evaluation was codified to direct the phlebotomist in the preoperative clinic or at any diagnostic center within our institution to draw the appropriate tubes of blood even if they are not familiar with this test. This order is available to all providers and all patients within our institution. The success of the automated reflex process involved extensive information technology innovations within the patient’s electronic record, the phlebotomy areas, the diagnostic laboratory receiving area, and the hematology and chemistry testing departments. Information technology processes had to direct phlebotomists to collect blood in the correct tubes and the routing mechanisms in the laboratory to send the blood tubes to the correct testing departments. Information technology support was necessary to trigger the chemistry laboratory to run the secondary tests based on the results of the complete blood count. The goal in developing the preprocedure anemia evaluation was to maximize efficiency and efficacy in diagnosing treatable types of anemia in preoperative patients. The time between when patients are scheduled for surgery and the surgery date is limited. Patients are inconvenienced by multiple appointments and often cannot return for another blood draw. The secondary goal was to decrease unnecessary testing with its associated costs. We could have obtained two vials of blood on every patient and ordered all of the tests regardless of the hemoglobin value but that would be wasteful for patients without anemia.

The preprocedure anemia evaluation is ordered in place of a complete blood count. The reflex tests to evaluate anemia are processed only when anemia is noted on the first component of the preprocedure anemia evaluation, the complete blood count (fig. 1). The only additional cost is the cost of the green-top blood collection tube containing heparin, which at our institution is $0.18. This cost is nominal compared with the previously described direct and indirect costs of anemia related to increased complications, length of stay, infection, and intensive care services. It is important to reiterate that anemia, like any other disease, should be treated. Addressing chronic conditions can improve both perioperative and long-term outcomes.

Diagnosis of iron-deficiency anemia is limited by an individual physician’s knowledge of the significance of anemia, expertise in evaluating anemia, and bandwidth for additional work.¹⁵  Providers often do not appreciate the significant association of adverse events in surgical patients with anemia³  and are often unaware of the importance of diagnosing iron-deficiency anemia. Physicians may have a sense of futility of managing patients with anemia in the preoperative period because they often think there is not time to evaluate and treat anemia preoperatively.

A limitation of this study is that we did not mandate use of the preprocedure anemia evaluation once it was implemented. As a result, despite extensive educational efforts, some preoperative clinic providers were slow adopters and continued to screen for anemia using the traditional approach. This was not a prospective study, so the patients screened for anemia after implementation were not randomly assigned to traditional anemia screening or the preprocedure anemia evaluation. Thus, there may be confounding factors that contributed to the increased proportion of screened patients diagnosed with iron-deficiency anemia by the preprocedure anemia evaluation. This study is also limited in that we could not account for patients who may have had iron studies outside of our institution. Our current practice uses the preprocedure anemia evaluation as the only diagnostic test for evaluating a preoperative complete blood count, diagnosing anemia, and testing for iron deficiency.

Diagnosing treatable causes of anemia in a timely manner preoperatively is an important component of patient blood management programs focused on limiting harmful and costly blood transfusions^8,9  Early identification of iron-deficiency anemia and treatment with intravenous iron infusions to optimize hemoglobin concentration preoperatively is recommended.^16,17 

This study shows that the use of the preprocedure anemia evaluation, a specially developed comprehensive testing process, is more effective and efficient in diagnosing iron-deficiency anemia than traditional methods of evaluating anemia in patients preoperatively. Early diagnosis should allow for earlier treatment of iron-deficiency anemia with a greater likelihood of correction of anemia. The scope of our study was not to evaluate outcomes or rates of transfusions. Exploring whether this expedited process for diagnosing anemia affects outcomes or need for transfusions are subjects of a future study.

The authors acknowledge Naveen Nathan, M.D., Northwestern University Feinberg School of Medicine, Chicago, Illinois, for his contribution to figure 1.

Supported by funding received from the Department of Anesthesiology, Northwestern University Feinberg School of Medicine, Chicago, Illinois.

Dr. Avram is on the editorial board of Anesthesiology, co-investigator on National Institutes of Health (Bethesda, Maryland) grant Nos. U54HD085601 (Wisner, George, Stika) and 1R01DC016645-01A1 SP0046825 (Tan), and is a co-mentor on grant No. 1K23HD087529-01A1 (Clark). Dr. Sweitzer is on the editorial board of Anesthesiology and receives compensation from the International Anesthesia Research Society (San Francisco, California).