The sternal brace: a novel osteopathic diagnostic screening tool to rule out cardiac chest pain in the emergency department

Methods

Test methods

Pretest and posttest NRS scores were obtained and compared with the final diagnosis. The final diagnosis was the final decision of the attending physician as written in the ED note for that visit. The adverse events from the reference standards were reported. Tests that were hindered or indeterminate were not included in the study.

Data analysis

The final diagnoses retrieved by R2 were then divided into two groups: CCP and non-CCP. Diagnostic test values were calculated utilizing MedCalc and Diagnostic Test Calculator [7, 8]. Various cutoff values of pretest/posttest decrease in chest pain were plotted for specificity and sensitivity as well as the receiver operating characteristic (ROC) curve that was generated and is shown in Table 1 and Figure 1, respectively. A decrease in chest pain of 2 was considered to be the best choice to study. Therefore, the test will be considered positive for cardiac etiology if the sternal brace does not reduce the NRS score by 2. A test will be considered negative if the sternal brace reduces the NRS by 2 or more. It may be counterintuitive to consider little to no decrease in chest pain as a positive test; therefore, Figure 2 was added to reiterate and clarify this idea.

Table 1:

The sensitivity and 1-specificity of each NRS test result based on its ability to differentiate CCP from non-CCP.

Sensitivity and specificity of various NRS decreases to predict cardiac issue
NRS change	1-Specificity (x)	Sensitivity (y)
1	0.57	0.73
2	0.65	0.82
3	0.74	0.82
4	0.83	0.82
5	0.87	0.91
6	0.91	1

1. CCP, cardiac chest pain; non-CCP, noncardiac chest pain; NRS, numerical rating scale.

Figure 1:

Receiver operating characteristic (ROC) curve plotting the sensitivity and specificity of each NRS score. A decrease in NRS of 2 was chosen to differentiate between cardiac and noncardiac.

Figure 2:

If the sternal brace has little effect on chest pain (i.e., NRS decreases by 0 or 1), then the test is considered positive and more likely to be cardiac. If the sternal brace has a good effect (i.e., NRS decreases by 2 or more), then the test is negative and less likely to be cardiac.

Cardiac ischemia in the US disease prevalence of 0.029 [9] was utilized for posttest probabilities of positive predictive value (PPV), negative predictive value (NPV), and accuracy Equations (1)–(3) of these posttest probabilities are displayed in Figure 3.

Figure 3:

Equations (1)–(3): PPV, NPV, and accuracy with prevalence included in each equation [7].

Results

A total of 34 patients were included in this study. Results obtained from the Newark Beth Israel ED in consenting adults between the age of 22 and 85 years old with a mean age of 48 ± 15 and a median age of 45.5. Race and gender data were not collected, yet demographics likely fell along the lines of Newark’s Weequahic neighborhood demographics: 46 % male, 54 % female; 91.9 % African American, 5.0 % Hispanic, and 3.1 % other races [10]. The clinical characteristics were not recorded except for age, medical record number, sternal brace test results, and final diagnosis.

To predict CCP, a cutoff value was a decrease in pain severity of 2 or greater between pretest and posttest NRS scores. A flow chart and diagnostic test table for this value are shown in Figure 4.

Figure 4:

(Left) A population size of 34 patients with unspecified chest pain, utilizing an NRS change of 2. Eleven patients had a final diagnosis of CCP, and 23 were noncardiac chest pain (non-CCP); there were 9 true positives, 2 false negatives, 15 false positives, and 8 true negatives. (Right) Diagnostic test table results for a cutoff value of 2.

Utilizing this cutoff value and prevalence yielded the following: Sensitivity 81.8 % (95 % confidence interval [CI], 48.2–97.7 %); specificity 34.8 % (95 % CI, 16.4–57.3 %); PPV 3.6 % (95 % CI, 0.1–20.3 %); and NPV of 98.4 % (95 % CI, 66.8–100.0 %). There was accuracy of 36.2 % (95 % CI, 20.4–54.4 %) [9]. There was a positive likelihood ratio of 1.25 (95 % CI, 0.8 to 1.9) with posterior probability (odds) at 37 % (95 % CI, 29–47 %); therefore, approximately 1 in 2.7 with a positive test are sick. We have a negative likelihood ratio of 0.52 (95 % CI, 0.13 to 2.06) with posterior probability (odds) at 20 % (95 % CI, 6–50 %); therefore, approximately 1 in 1.2 with a negative test are well [9]. A full list of compiled diagnostic test values is shown below in Table 2.

Discussion

The physiologic basis and mechanism of the sternal brace is not known; however, the authors theorize that the maneuver relieves the dysfunction of the innermost/internal intercostal muscle. These muscles were chosen because they are located both medially and laterally in the chest wall. Similarly in this study, chest pain location could be located anywhere along the midline sternum or as far lateral as the axillary line. It is less likely to be the transversus thoracis because it does not extend laterally, or the external intercostals, which are not located medially.

The innermost/internal intercostals are muscles for expiration. Chest pain is often exacerbated by inhalation, which stretches the innermost/internal intercostals [5]. When the maneuver compresses the chest wall, then the chest wall is put into an exhalation position and the length of the innermost/internal intercostals is reduced. Then, the patient’s attempt to actively expand the rib cage makes the innermost/internal intercostals reciprocally inhibited and not allowed to lengthen because of the operator’s opposing force keeping their length shortened. Therefore, this could then be considered a counterstain technique or a facilitated position-release technique.

The sternal brace study is important for a number of reasons.

1. There are few physical examination findings to separate CCP from the more common causes of chest pain. Therefore, the sternal brace adds a valuable perspective to the diagnosis.

2. The high NPV lends itself to triaging patients in a busy ED or with emergency medical services in the field.

3. The examination is quick, painless, and easy enough that it can be taught to many different levels of healthcare workers.

4. Chest pain can be psychologically distressing for patients and detrimental to their mental health, which is a reassuring sign of pain relief could be beneficial.

5. This could be a helpful screening tool when more advanced equipment is not readily available.

Finally, this is an osteopathic technique that has been adapted into a diagnostic physical examination finding. There are many such osteopathic techniques that are underutilized by the medical profession as a whole that may provide better therapeutic and diagnostic techniques.

Study limitations including sources of potential bias, statistical uncertainty, and generalizability are reported here. The demographic of Newark, New Jersey is not representative of the United States as a whole. We had a small sample size of 34 patients because of limited time availability for recruitment. Because this was a convenience sample, a wide age range was recruited, which may confound the data because CCP is highly unlikely in the younger cohort. Race and gender were not collected because this was a proof-of-concept pilot study. Determining CCP was purely based on the final diagnosis of the ED attending during that ED visit without the consideration of laboratory work, other testing, and further hospital course; these would have been valuable to corroborate the final diagnosis. CCP possibly had a higher prevalence in this study because those patients highly suspicious for cardiac issues tend to have a longer ED course and therefore a higher chance of the study team recruiting them. Additionally, more benign forms of chest pain, including musculoskeletal chest wall syndromes and gastroesophageal reflux disease (GERD), may be underrepresented; presumably these more benign forms of chest pain would be discharged more quickly and therefore are less likely to be recruited in this convenience sample. However, it is possible that those patients with more severe symptoms were less amenable to consenting to an experimental study in the first place.

In future studies, we recommend: (1) a more thorough corroboration of the final diagnosis in the ED visit; (2) an increase in sample size; (3) a cohort study of the effect of the sternal brace on known CCP patients; (4) utility of the sternal brace in triaging patients with a chief complaint of chest pain; and (5) better categorization of non-CCP.

Conclusions

The primary objective was to assess the sternal brace as a screening tool. Therefore, with a sensitivity of 81.8 % and an NPV of 98.5 %, this test would be a good, quick screening tool to rule out CCP. Therefore, if a person with chest pain has an NRS that decreases by 2 or more with this maneuver, then there is a 98.5 % chance that the chest pain is noncardiac, given its prevalence. Also, if the disease is present, then it is 81.8 % likely that their NRS will not decrease by more than 2. This test did not perform well in other diagnostic measures; it had poor specificity, poor PPV, and poor accuracy.