Cost of treating peripheral neuropathic pain with pregabalin or gabapentin at therapeutic doses in routine practice

Tha database used for this secondary analysis was formed by 86,206 electronic medical records of patients over 18 years of age. A total of 923 records of patients were recruited who met the inclusion/exclusion criteria described previously. There were 711 patients (77.0%) in the pregabalin study group and 212 patients (23.0%) in the gabapentin study group. Mean age was 59.4 (standard deviation: 14.6) years and 62.3% were female. Table 1 shows the baseline characteristics of patients analyzed according with the therapeutic group. Also, it shows the distribution of the different types of PNP by study group. Numerical differences were observed between age and gender, although these were not statistically significant: a mean age of 59.8 years on pregabalin versus 58.0 years with gabapentin (p = 0.108); and a proportion of females of 64.1% versus 56.1% (p = 0.042). The proportion of associated comorbidities, RUB (3.1 vs 3.2; p = 0.321) and Charlson index (0.5 vs 0.6; p = 0.177) were similar for both groups studied. PNP was caused by some sort of painful focal neuropathy in near 65% of patients studies, with about 45% of them being of axial radiculopathy origen (cervical or lumbar). Polineuropathy was diagnosed in 21.5% of patients, with diabetic polyneuropathy being the most prevalent in the database. A total of 12% of cases were due to trigeminal/postherpetic neuralgia. The distribution of PNP showed no statistically significant differences regarding the use of pregabalin or gabapentin. Table 2 shows the characteristics of medication use. In general, mean duration of treatment with pregabalin was slightly lower than with gabapentin (5.4 vs 5.8 months), although without reaching statistical significance. Only 5.7% of patients in the gabapentin group were receiving doses above 1800 mg per day, while 11.5% of pregabalin subjects were treated with 600 or more mg per day.

Table 3 shows the gross and adjusted costs (24-month follow-up) associated with PNP according to study group. Healthcare costs accounted by 53.7% of the total costs, while 46.3% corresponded to productivity losses due to days of sick leaves. The average per patient cost was €2653.40. Of the total costs, 40.2% were incurred in primary care and 13.5% were located in specialized care. Of the latter, 17.3% were incurred in drug prescription. The average cost/patient for antiepileptic medication was similar; €229.20 for pregabalin versus €224.20 for gabapentin (p = 0.762), although the cost of concomitant analgesic medication was significantly lower for pregabalin; €184.10 versus €386.20 (p < 0.001). By group, the total costs (healthcare and non-healthcare) for the patients on treatment with pregabalin were lower than with gabapentin (€2479.00 vs €3238.20; p = 0.006, Table 3). These differences remained after adjusting for co-variates. The mean adjusted total cost per patient was significantly lower with pregabalin than with gabapentin; €2463.60 for pregabalin versus €3142.10 for gabapentin (p = 0.014), with a mean adjusted difference of €678.50 in total costs (p = 0.014). The component of cost causing this significant difference was the one located at the primary care level (€298.80 on average per subjects; p < 0.001). The majority of the cost components were lower with pregabalin, except for those observed in specialized care, that were not statistically different. Table 4 lists the distribution of the main cost components according to gender, age and main diagnoses by study group. It should be noted that these differences were maintained in all subgroups, such that the patients on treatment with pregabalin were associated with a lower total cost. In all these subgroups, total costs were lower with pregabalin and the component explaining the main difference was again the one located at the primary healthcare level. The differences in postherpetic/trigeminal neuralgia and diabetic painful polineuropathy subgroups were not statistically different because of the small size of such groups.

This study sets out the cost of PNP treatment with either pregabalin or gabapentin within the range of therapeutic doses, given that this is relevant information for healthcare decision-making in routine clinical practice in our healthcare setting. There have been very few economic studies conducted in Spain or other countries comparing the cost of PNP treatment with pregabalin and gabapentin and we are not aware of any with our study's objective. However, it should be noted that without appropriate standardization of patient characteristics, as well as in the number and extent of the variables studied, the results obtained must be interpreted with caution, and care should be taken in the external validation of the results.

Without taking into account the therapeutic range of the drug (pregabalin/gabapentin) [10], the total cost of PNP per patient was around €728 lower in those treated with pregabalin than in those treated with gabapentin during the 2 years of patient follow-up, €162 of which was for the healthcare cost component [10]. This difference was explained by lower use of healthcare resources (mainly medical visits and concomitant analgesic medication) in the patients who received pregabalin in comparison with gabapentin, which clearly offset the greater cost of procuring pregabalin versus gabapentin [10]. When the effect on patients who received doses in the therapeutic range is analyzed, in whom the difference between pregabalin and gabapentin in the cost of antiepileptic medication is also no longer apparent, the differences between the two drugs persist. In other words, there is a notable saving with pregabalin in comparison to gabapentin (difference: €679/patient), again due to the lower use of concomitant analgesic medication and fewer medical visits. These results continue to be consistent when analyzed separately by gender, age group and different types of aetiology causing PNP, except for some neuralgias in which the differences did not reach statistical significance owing to the small number of patients. One of the reasons that may explain such findings is that titration until therapeutic dosage is faster (or more ease) with pregabalin in comparison with gabapentin. Thus, subjects receiving gabapentin need more time to reach a dose within the range of its recommended therapeutic dosage, and also due to its particular formulation and pharmacodynamics (need be taken three-times a day) many patients have difficulties in receiving the appropriate dosage under routine medical care, that differs from what occurred in traditional randomized clinical trials. As a consequence, these patients use more visits to physicians and more concomitant analgesics than with pregabalin, and also days of sick leave although this component was not statistically different between drugs. Also, we cannot rule out completely that patients who are treated with gabapentin under routine medical practice are difficult to be treated to higher dosages such as the one observed in randomized clinical trial with such drug.

The results observed in our retrospective study are consistent with other series published in the scientific literature either in Spain [19,20], and elsewhere as well [21–24]. All of such references, irrespectively of the economic type of study (cohort observational trial, economic modeling or clinical trial) have shown that in different types of PNP, pregabalin therapy is normally associated with lower healthcare resources utilization and fewer days of sick leave in comparison with gabapentin leading to a lower economic impact to the National Health System budgets, which is relevant when health decision is taken on a routine basis. As commented previously, physicians in their routine practice should be aware that titration with gabapentin is more difficult than with a more novel drug such as pregabalin, abiliting patient to reach the therapeutic analgesic dosage no only faster, but an effective doses. Two more recent publications, from Wang et al. [25] and from Schaefer and coworkers [26], concludes that pregabalin is an effective treatment for postherpetic neuralgia and PNP in general, the higher cost of the drug being offset by lower pain scores and more pain-free days, and corroborates the reduction of work-related costs in patients treated with pregabalin. Despite variations in methodology, these studies are consistent with our results. Part of the difference in costs between pregabalin and gabapentin was attributed to the costs deriving from primary care, while the differences observed in secondary care did not reach statistical significance. Again, findings that primary care physicians should be awarded of difficulties when titration with gabapentin in the real word. Thus, the information that this study provides may be important not only for physicians but also for healthcare decision-makers, particularly those interested in the area of PNP.

As this study was designed as a retrospective one using existing data, we certanly have to deal with potential limitations, that could include errors in disease categorization (type of PNP), under-recording of the disease, potential variability of professionals and patients and classification bias and choice of therapeutic groups selected by the prescribing physician, the so called selection bias. In the statistical analysis, no important differences were observed in the comparability of the groups when starting treatment with the drugs assessed in the study, at least in the variables analyzed, such as demographic characteristics and associated comorbidities. Nonetheless, the analysis tried to control for confounding co-variates, such as gender, age, time since diagnosis, Charlson index and the resources utilization band (RUB), that were considered to have a potential role on the main end points of the study. Also, to gain in comparability, the study extracted records with exposition to the drugs analyzed for the first time. Finally, but not less important, the study was not able to obtain measurement of pain severity and treatment adherence. Likewise, the tolerability profile (adverse event) of the drugs. The last because in the database, potential side effects of drugs are captured in a narrative way in free open fields in the database, being difficult the statistical analtysis of such data.

Future research must involve studies of cost–effectiveness and diagnostic and treatment delay, in addition to replicating the study in other healthcare organizations (generalization of the results). Given that, owing to the study design, it cannot be known whether pregabalin has a better safety profile or better pharmacokinetic properties, randomized clinical trials should be conducted that compare the results obtained in this study.

Despite the limitations indicated, on a population level and in a routine clinical practice setting, when prescribed at doses in the therapeutic range recommended for the treatment of PNP, pregabalin was associated with lower healthcare and non-healthcare costs in comparison with gabapentin, particularly those costs relating to analgesic drugs and medical visits. This results in a lower total cost for the Spanish National Health System.

This article has shown that treating peripheral neuropathic pain with pregabalin was associated with lower costs to the National Health System in Spain in comparison with gabapentin. As a consequence, these findings may be of help to payers and health decision makers who are in charge of healthcare budgets. On the other hand, this research postulated that pharmacodynamics differences of pain medications evaluated here, could be one of the reasons explaining the findings of the study. Thus, utilization of pain drugs with easy-to-take formulations could be associated with less healthcare resources utilization and analgesics consumption.

The study was conceived and designed by A Sicras-Mainar, J Rejas-Gutiérrez and María Pérez-Páramo, data collection and statistical analysis were done by ClinicResearch. R Navarro-Artieda was responsible for literature review and bibliographic support. All authors contributed to the data interpretation, drafting, revision and approval of the submitted manuscript.

Data collection and statistical analysis were carried out by ClinicResearch, a Healthcare Services research Network, and was funded by Pfizer, SLU. English editing of manuscript was carried out by Welocalize Life Science and was funded by Pfizer, SLU. J Rejas-Gutiérrez and M Pérez-Páramo are employees of Pfizer SLU. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

The Authors state that the study has been classified by the Spanish Agency for Medicines and Medical Devices (AEMPS) as a Post-Authorization Study – Other Design (EPA-OD) and subsequently approved by the Independent Ethics Committee of Hospital Universitari Germans Trias i Pujol in Badalona. In addition, the confidentiality of the records established by the Spanish Organic Data Protection Law (15/1999, of 13 December) has been respected, with dissociation of the data.

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/