Risk prediction and impaired tactile sensory perception among cancer patients during chemotherapy 1

ABSTRACT Objectives: to estimate the prevalence of impaired tactile sensory perception, identify risk factors, and establish a risk prediction model among adult patients receiving antineoplastic chemotherapy. Method: historical cohort study based on information obtained from the medical files of 127 patients cared for in the cancer unit of a private hospital in a city in Minas Gerais, Brazil. Data were analyzed using descriptive and bivariate statistics, with survival and multivariate analysis by Cox regression. Results: 57% of the 127 patients included in the study developed impaired tactile sensory perception. The independent variables that caused significant impact, together with time elapsed from the beginning of treatment up to the onset of the condition, were: bone, hepatic and regional lymph node metastases; alcoholism; palliative chemotherapy; and discomfort in lower limbs. Conclusion: impaired tactile sensory perception was common among adult patients during chemotherapy, indicating the need to implement interventions designed for early identification and treatment of this condition.


Method
This is a historical cohort study with analysis of the data contained in the medical files of patients cared In the study period, information contained in the files of 253 patients admitted and accessed in the aforementioned oncological unit was analyzed. Of these, 103 did not meet the inclusion criteria, while the files of five patients were not located, so that a final sample of 127 patients was obtained.
The participants had their files analyzed at the time of admission, at every ACH cycle, that is, every month, and at the end of the ACH or when the condition emerged. Two instruments were used to collect data.
The first, filled out at the time of admission, addressed sociodemographic and clinical data, prior oncological treatments, and symptomatology of impaired tactile sensory perception. The second instrument was meant to record subsequent assessments that corresponded with the patients' answers for every ACH cycle, and addressed clinical data and risk factors for the development of this health condition (3,(5)(6)(7)(8)(15)(16) . Note that information was based on medical and nursing reports contained in the patients' files.
Note that the translated version of the AINQ validated for Brazilian Portuguese (11)  Variables with a p-value ≤0.25 in the bivariate analysis were included in the multivariate analysis model. The global adjustment of the model was obtained by means of the likelihood ratio test (18) .
This study complies with Resolution 466/12, which regulates research with human subjects, while the project was approved by the Institutional Review Board (CAAE -192666113.7.0000.5149).

Results
Among the 127 medical files assessed, 73 presented altered sensory perception. The global incidence was 57% in the study period, corresponding to an incidence rate of three cases for each 1,000 patients, ranging from two to four, with a confidence interval of 95%.
Of the total patients, 77% originated from oncological clinics, had a diagnosis of gastrointestinal tract cancer (28.3%), hematological (24.4%), gynecological (22.8%), or lung malignant disorders (12.6%), or disorders in other body systems (9.4%), in addition to unknown primary sites (2.5%). At the time of admission, we verified that 51% of the patients presented distant metastases, that is, had stage IV cancer. After this level, ACH is defined as palliative care (2) . The most common sites of metastasis were the liver (13%), regional lymph nodes (11%), lungs (10%), and peritoneum (9%). Most (80%) patients were receiving the ACH for the first time and the surgeries most frequently reported as prior treatments were lymphadenectomy (57%) and tumor resection (47%).
The bivariate analysis presented variables that were statistically (p≤0.25) associated with time since the beginning of treatment up to the onset of impaired tactile sensory perception. A total of 80 variables were eligible for the multivariate analysis, 33 of which were statistically significant (p<0.05) and are presented in Table 1.
A predictive risk model for tactile sensory perception among adult patients receiving ACH resulted from the multivariate analysis ( Table 2). Note that among the factors identified, bone, hepatic and lymph regional node metastasis, in addition to palliative chemotherapy, and discomfort in the lower limbs, presented significant

Discussion
Impaired tactile sensory perception is a subjective problem, that is, patients need to communicate their symptoms, which should be highlighted as a recommendation when care is provided to cancer patients receiving neurotoxic ACH (11) . A total of 57% of the population under study presented altered tactile sensory perception and reported physical weariness due to the treatment including paresthesia in both hands and feet, discomfort in the upper limbs, lack of or decreased touch sensitivity and hypoesthesia, situations widely reported in the literature (5)(6)(7)(8)(9)11) . The rates of impaired tactile sensory perception reported in the literature range from 51% and 98% (6)(7)(10)(11) , depending on the type of medication involved.
The variables presented in this study as factors that predispose cancer patients receiving ACH to a lower risk for the development of the outcome under study were: palliative ACH; peripheral venous access to administer ACH; sixth chemotherapy cycle; and the use of carboplatin. These findings diverge from those reported in the literature, which suggest a relationship between the onset of impaired tactile sensory perception in patients receiving palliative ACH such as paresthesia and dysesthesia, and the cumulative effect of the medications used (6,9,11) . Among the variables that presented significant association (p<0.05) with time elapsed from the beginning of the treatment and the onset of impaired tactile sensory perception, and which predispose patients to a greater risk of developing the outcome under study (HR≥1), researchers (5)(6)(7)(8)(9)(10)(11)19) report the use of Oxaliplatin, the emergence of Peripheral Sensory Neuropathy (PSN), report discomfort in lower and upper limbs when in contact with ice, paresthesia, dysesthesia, peripheral discomfort, lack of/decreased touch and sensitivity, and difficulty in performing manual tasks.
Note that the use of Oxaliplatin (HR=1.63, p=0.049) in this study is similar to third generation platinum, commonly used with patients with colon cancer. It promotes neurotoxicity, leading to important loss or altered sensory situations such as decreased touch perception (9) . In this sense, nurses should be attentive to include an assessment of the integrity of patients' sensory function in the nursing care plan of those receiving Oxaliplatin (20) .
ACH induced sensory neuropathy, especially degree one neuropathy, was identified as one of the main risk factors for impaired tactile sensory perception (HR=6.45, p<0.001). One recent study (10) , however, reports that deficient tactile sensation was a predictor of chemotherapy-induced peripheral neuropathy in patients who received Oxaliplatin to treat colorectal cancer.
In this study, complaints concerning lack of or decreased touch sensitivity manifested by the patients emerged as important risk factors to be identified whenever impaired tactile sensory perception is identified, as they reflect damage caused to structures essential for receiving and transporting electrical impulses, which can be caused by ACH. The interruption of electrical impulses is the main characteristic of ACHinduced sensory neuropathy, toxicity basically caused by axonopathy (5)(6)9) . These results indicate that patients receiving neurotoxic ACH often experience ACH-induced sensory neuropathy and that such a tactile disturbance is the individual's response to such toxicity.
In regard to carboplatin, the results are similar to those presented in the literature. This platinum is known for having lower neurotoxicity potential (21) . Neurological symptoms only occur when ACH is administered in high doses, in elderly patients, associated with taxon, or in individuals with prior exposure to another drug with neurotoxic effects (3) .
The multivariate analysis revealed six co-variables that were statistically significant (p<0.05) for time since